Undergraduate BIOL201 ⭐ 3 Credits Semester 4

Molecular Biology: Mechanisms of Gene Expression

In-depth examination of the molecular mechanisms of DNA replication, transcription, RNA processing, and translation. Emphasizes biochemical mechanisms, structural biology of the protein and nucleic acid machines involved, and the regulation of gene expression at each step. Essential for understanding the cellular machinery that produces proteins.

📅 Studied by Skipper on 2026-10-01

📖 Core Concepts

Claude Opus 5.5

BIOL201: Core Concepts. Mechanisms of Gene Expression

1. What Is This Subject Really About?

Most textbooks open this subject with a diagram of the Central Dogma: DNA → RNA → protein. It is correct, but it hides almost everything interesting. Molecular biology is less a story about information than a story about machines that handle information. These machines copy it, read it, edit it, check it, and decide when to act on it. Each one is built from the same few chemical ideas you already know: nucleophilic attack on phosphorus and carbonyl carbon, the coupling of unfavorable steps to favorable ones, and molecular recognition through hydrogen bonding and shape complementarity.

Here is a frame that will serve you well. Each step of gene expression must solve four problems:

  1. Specificity. How does the machine pick the right substrate from a crowd of nearly identical ones? Examples are the correct nucleotide among four, the correct tRNA among roughly forty, and the correct splice site among thousands of look-alikes.
  2. Energy. Where does the free energy come from to build an ordered polymer, and how is it spent?
  3. Directionality and processivity. How does the machine move one way, stay on its template, and know where to start and stop?
  4. Regulation. How does the cell decide whether, when, and how much?

If you can answer these four questions for replication, transcription, RNA processing, translation, and protein targeting, you understand molecular biology rather than having memorized it. The rest of these notes are organized around that idea.

"The central dogma... deals with the detailed residue-by-residue transfer of sequential information. It states that such information cannot be transferred back from protein to either protein or nucleic acid."
— Francis Crick, [Crick, 1970]

Note what Crick actually said. The dogma does not forbid RNA → DNA transfer (reverse transcriptases do exactly that, telomerase among them). It forbids information flowing out of protein back into sequence. Keep this in mind for the misconceptions section.

A second recurring theme, and probably the most surprising one in the course, is that RNA does the chemistry at the two most ancient steps. Peptide bond formation and pre-mRNA splicing are catalyzed at their cores by RNA, with proteins acting as scaffolds, chaperones, and motors. This is widely read as molecular fossil evidence of an "RNA world" that preceded DNA and protein.


2. Template-Directed Polymerization: One Reaction, Three Machines

DNA polymerases, RNA polymerases, reverse transcriptases, and telomerase all catalyze the same chemical reaction. Learn it once and you have learned it four times.

Nucleotidyl transfer (phosphodiester bond formation):

  (NMP)n–3'-OH  +  NTP  ──Mg²⁺──▶  (NMP)n+1–3'-OH  +  PPi

  Mechanism: the 3'-OH of the primer terminus (activated by Mg²⁺ "A")
  attacks the α-phosphorus of the incoming NTP (whose triphosphate
  is coordinated by Mg²⁺ "B"); in-line SN2-type displacement at P;
  pyrophosphate is the leaving group.

Coupled reaction that drives polymerization forward:

  PPi  +  H₂O  ──pyrophosphatase──▶  2 Pi      ΔG°' ≈ −19 kJ/mol

This is the two-metal-ion mechanism [Steitz, 1998]. Connect it to CHEM202: it is nucleophilic substitution at phosphorus, the phosphate analog of acyl substitution at a carbonyl. One Mg²⁺ lowers the pKa of the 3'-OH so it becomes a better nucleophile. The other stabilizes the developing negative charge on the pentacovalent transition state and on the departing pyrophosphate. The polymerase is in essence a precisely shaped pocket that positions two metal ions.

Why always 5'→3'? A design argument

Every known nucleic acid polymerase extends chains 5'→3'. This is not an accident. It follows from proofreading:

  • In 5'→3' synthesis, the high-energy triphosphate is carried by the incoming nucleotide. If a wrong nucleotide is added and then excised by a 3'→5' exonuclease, the chain end is a plain 3'-OH, ready for the next correct dNTP. Nothing is lost.
  • In a hypothetical 3'→5' synthesis, the triphosphate would sit on the growing chain's 5' end. Excising a mismatched terminal nucleotide would remove that activating group and leave a dead chain that would need a separate reactivation step.

The directionality of life's polymers is thus a consequence of the need to correct errors cheaply [Alberts et al., 2022]. When you meet a "rule" in biology, ask whether it is a frozen accident or a design constraint. This one is a constraint.

The polymerase cannot start from nothing

DNA polymerases need a 3'-OH to extend. They cannot initiate de novo. This one fact generates a cascade of consequences: RNA primers made by primase, Okazaki fragments on the lagging strand [Okazaki et al., 1968], and the end-replication problem at linear chromosome tips. RNA polymerases can initiate de novo. That is part of why transcription can begin anywhere a promoter says it should.


3. DNA Replication: Fidelity, Licensing, and the Problem of Ends

3.1 The semi-conservative proof

Meselson and Stahl grew E. coli in 15N medium, shifted the cells to 14N, and separated the DNA by equilibrium density-gradient centrifugation in CsCl [Meselson & Stahl, 1958]. After one generation all DNA was of intermediate density, which ruled out conservative replication. After two generations there were equal intermediate and light bands, which ruled out dispersive replication. This connects directly to PHYS201: DNA molecules come to rest where their buoyant density equals the local density of the CsCl gradient that forms under centrifugal force. It is also one of the most elegant experiments in biology because each model made a distinct, quantitative, falsifiable prediction.

3.2 Fidelity is multiplicative

Watson–Crick pairing alone discriminates only roughly 1 in 102–103, because the free-energy difference between a correct and an incorrect base pair in solution is small. The cell reaches about 1 error per 109–1010 nucleotides by stacking independent checks:

Overall error rate ≈ (base selection) × (proofreading) × (mismatch repair)

                   ≈ (10⁻⁴ to 10⁻⁵) × (10⁻²) × (10⁻² to 10⁻³)

                   ≈ 10⁻⁹ to 10⁻¹⁰ per nucleotide per replication

Base selection beats solution chemistry because the polymerase active site checks geometry. Correct pairs share nearly identical C1'–C1' distances and minor-groove hydrogen-bond acceptor positions, and the "fingers" domain closes only around a correctly shaped pair (induced fit). The same principle, that shape is checked in the minor groove, reappears at the ribosome's decoding center. Watch for it.

3.3 Bacterial machinery

  • DnaA (an AAA+ ATPase) binds 9-mer "DnaA boxes" at oriC in its ATP-bound form, oligomerizes, and melts the AT-rich 13-mer region. AT pairs have two hydrogen bonds and weaker stacking, so they melt more easily than GC.
  • DnaB helicase is loaded by DnaC and translocates 5'→3' on the lagging-strand template. SSB coats the exposed single strands to prevent re-annealing and hairpins.
  • DNA Pol III holoenzyme is a dimeric (in fact often trimeric) core tethered by the clamp loader (τ/γ complex). The β-clamp is a ring that encircles DNA and raises processivity from about 10 nucleotides to many thousands. Speed is about 1000 nt/s.
  • Termination: Tus protein bound to Ter sites acts as a polar "one-way gate" that lets forks enter the terminus region but not leave it.

Analogy: the β-clamp is a carabiner clipped around a rope. The polymerase can slide fast without falling off, and the clamp loader is the hand that opens and closes the carabiner, spending ATP to do so.

3.4 Eukaryotic replication: once and only once

A human genome has tens of thousands of origins. If any fired twice in one S phase, segments would be over-replicated, a recipe for genomic instability. The solution is a two-step licensing system that separates loading from activation in time [Watson et al., 2014]:

  1. G1 (low CDK activity): licensing. ORC binds origins and recruits Cdc6 and Cdt1, which load the MCM2–7 helicase as an inactive head-to-head double hexamer around double-stranded DNA.
  2. S phase (high CDK and DDK activity): firing. CDK and DDK (Cdc7–Dbf4) phosphorylate their targets, recruiting Cdc45 and GINS to form the active CMG helicase (Cdc45–MCM–GINS), which unwinds DNA.
  3. The same high CDK activity blocks new licensing by phosphorylating and inactivating Cdc6, ORC, and (in metazoans) Cdt1, with geminin also inhibiting Cdt1. New helicases cannot be loaded until CDK activity drops at the end of mitosis.

The beauty of this design is that one signal, CDK activity, both permits firing and forbids reloading. No origin can fire twice in one cycle because it cannot be relicensed until the cycle resets. Replication timing (early vs. late origins) correlates with chromatin state: euchromatic, transcriptionally active regions tend to replicate early, heterochromatin late.

3.5 Telomeres and telomerase

On a linear chromosome, the RNA primer at the extreme 3' end of the lagging strand cannot be replaced with DNA, because no upstream 3'-OH exists to extend. The chromosome would shorten every division. Telomerase solves this [Greider & Blackburn, 1985]. It is a reverse transcriptase (TERT) that carries its own RNA template (TR/TERC). It extends the 3' overhang by repeated cycles of TTAGGG addition in humans, translocating along its internal template between repeats. Conventional lagging-strand synthesis then fills in the complementary strand.

Telomerase is RNA-templated DNA synthesis, the reverse of transcription, operating inside every proliferating human cell. That is why Crick's dogma must be read carefully.


4. Bacterial Transcription: Sigma, Promoters, and Two Ways to Stop

Bacterial RNA polymerase core (α2ββ'ω) can transcribe but cannot find promoters. Adding a σ factor forms the holoenzyme, which recognizes promoter elements centered near −10 (consensus TATAAT) and −35 (TTGACA). The spacing between them (17 ± 1 bp optimal) matters because the two sites must sit on the same face of the polymerase.

The initiation pathway

R + P  ⇌  RPc  ⇌  RPo  →  initial transcribing complex  →  elongation complex
       (closed)  (open, ~13 bp melted,      (abortive cycling,       (σ released,
                  −11 to +2)                 "scrunching")            processive)
  • Closed → open: σ region 2 flips out bases of the non-template strand in the −10 element and captures them in protein pockets. No ATP is needed; the binding energy of the holoenzyme pays for melting.
  • Abortive initiation and scrunching: the polymerase stays fixed on the promoter while it pulls downstream DNA into itself, storing strain like a compressed spring [Kapanidis et al., 2006]. Many short transcripts (2–15 nt) are released before the stored energy is enough to break σ–promoter contacts and launch promoter escape.
  • Alternative σ factors (e.g., σ32 for heat shock, σ54 for nitrogen metabolism) redirect the same core to new gene sets. This is regulation by swapping the "reading glasses."

Termination

  1. Intrinsic (Rho-independent): a GC-rich inverted repeat forms an RNA hairpin, followed by a run of U's in the transcript. The hairpin invades the RNA exit channel and destabilizes the complex. The rU:dA hybrid is the weakest of all base-pair combinations, so the remaining RNA–DNA hybrid falls apart.
  2. Rho-dependent: Rho is a hexameric, ring-shaped, ATP-dependent 5'→3' RNA translocase/helicase. It loads on C-rich, unstructured rut sites, chases the polymerase, and dislodges it when the polymerase pauses. Because translating ribosomes normally cover the nascent mRNA in bacteria, Rho mainly acts where translation has stopped. This is a built-in coupling of transcription to translation.

5. Eukaryotic Transcription and Chromatin: Regulation at a Distance

5.1 The Pol II preinitiation complex

Eukaryotes use three nuclear RNA polymerases: Pol I for most rRNA, Pol II for mRNA and most miRNA precursors, and Pol III for tRNA, 5S rRNA, and U6 snRNA. Pol II cannot recognize promoters alone. The general transcription factors assemble a preinitiation complex (PIC):

  • TFIID: TBP plus TAFs. TBP binds the TATA box in the minor groove and bends DNA by about 80°, a striking structural distortion.
  • TFIIB: bridges TBP and Pol II and helps set the start site.
  • TFIIF: escorts Pol II.
  • TFIIE: recruits TFIIH.
  • TFIIH: contains XPB, an ATP-dependent DNA translocase that opens the promoter (unlike bacteria, Pol II promoter opening requires ATP), and CDK7, which phosphorylates the Pol II CTD.

5.2 The CTD: a programmable landing pad

The C-terminal domain of the Pol II largest subunit has 52 tandem repeats (in humans) of the heptapeptide Y1S2P3T4S5P6S7. Its phosphorylation pattern changes as Pol II moves along the gene:

  • Ser5-P (by CDK7) is high near the promoter. It recruits capping enzymes.
  • Ser2-P (by CDK9/P-TEFb) rises toward the 3' end. It recruits splicing and 3'-end processing factors.

This is the physical basis of co-transcriptional processing. The CTD works like a tool belt whose contents change as the worker walks down the assembly line. A capped, spliced, polyadenylated mRNA is largely made while it is transcribed, not afterward.

5.3 Mediator, enhancers, and insulators

Enhancers are DNA elements that bind activator proteins and can stimulate transcription from thousands to hundreds of thousands of base pairs away, upstream, downstream, or within introns, in either orientation. They act through DNA looping, which brings enhancer-bound activators into contact with the promoter. The Mediator complex (about 25–30 subunits) is the molecular interpreter. It binds activation domains on one side and Pol II on the other, and it relays the combined input of many activators into PIC assembly and Pol II activity.

Insulators such as CTCF-binding sites solve the problem this creates: if enhancers act at a distance, what stops them from activating the wrong gene? CTCF, together with cohesin, defines loop domains (topologically associating domains, TADs) and restricts which promoters an enhancer can reach.

5.4 Chromatin: the default state is "off"

A nucleosome wraps 147 bp of DNA in about 1.65 left-handed superhelical turns around a histone octamer (two each of H2A, H2B, H3, H4) [Luger et al., 1997]. This packaging is repressive by default, since it occludes binding sites. Eukaryotic gene regulation is therefore largely about overcoming chromatin, whereas bacterial regulation is largely about recruiting polymerase to naked DNA.

Three classes of tools modify chromatin:

  1. Covalent histone modifications on the histone tails:
    • Lysine acetylation (by HATs; removed by HDACs) converts the positively charged ε-NH3+ into a neutral amide. Your organic chemistry predicts the effect: weaker electrostatic attraction to the DNA phosphate backbone and to neighboring nucleosomes. More important, acetyl-lysine is read by bromodomains found in activators and remodelers.
    • Lysine methylation (mono-, di-, tri-) does not change the charge, so its effects come entirely from readers such as chromodomains and PHD fingers. Context determines meaning: H3K4me3 marks active promoters, H3K36me3 marks transcribed gene bodies, H3K27me3 (Polycomb) marks facultative repression, and H3K9me3 (bound by HP1) marks constitutive heterochromatin.
  2. ATP-dependent remodelers (SWI/SNF, ISWI, CHD, INO80 families) use ATP hydrolysis to slide, evict, or restructure nucleosomes. They act as DNA translocases that pump DNA around the octamer.
  3. Histone variants such as H2A.Z at promoters and CENP-A at centromeres change nucleosome properties locally.

Writers, readers, and erasers form a logic system often called the "histone code." Treat the term with care. It is less a rigid code than a combinatorial set of binding platforms whose interpretation depends on context.


6. Pre-mRNA Processing: Capping, Polyadenylation, and Splicing

6.1 The 5' cap: an unusual bond

The cap is 7-methylguanosine linked 5'-to-5' through a triphosphate bridge to the first transcribed nucleotide (m7GpppN). It is a "backwards" nucleotide, which makes it invisible to 5'→3' exonucleases. It is installed in three steps on Ser5-phosphorylated CTD, when the transcript is only about 20–30 nt long:

1. RNA triphosphatase:    pppN–RNA  →  ppN–RNA  +  Pi
2. Guanylyltransferase:   GTP  +  ppN–RNA  →  GpppN–RNA  +  PPi
                          (via a covalent enzyme–lysine–GMP intermediate)
3. Guanine-N7 methyltransferase:
                          GpppN–RNA  +  SAM  →  m⁷GpppN–RNA  +  SAH

Note step 3: S-adenosylmethionine (SAM) is the cell's universal methyl donor, an SN2 transfer from a sulfonium ion. Higher eukaryotes also add 2'-O-methyls to the first one or two nucleotides ("cap 1/cap 2"). This helps innate immunity distinguish "self" mRNA from viral RNA. The cap later recruits CBC (cap-binding complex) in the nucleus and eIF4E in the cytoplasm.

6.2 Cleavage and polyadenylation

The 3' end of most eukaryotic mRNAs is not where transcription stops. It is made by cutting the transcript. CPSF recognizes the AAUAAA hexamer, and CstF binds a GU/U-rich element downstream. Cleavage occurs about 10–30 nt downstream of AAUAAA (CPSF73 is the endonuclease). Poly(A) polymerase then adds about 200–250 adenosines in mammals, a template-independent reaction, and PABPN1 coats the tail. The downstream RNA fragment still attached to Pol II is degraded by the 5'→3' exonuclease Xrn2, which helps dislodge Pol II. This is the "torpedo" model of termination.

6.3 Splicing: two transesterifications

This is one of the most important mechanisms in the course. Learn it at the level of individual atoms.

Step 1 (branching):
  2'-OH of branch-point adenosine  ──attacks──▶  phosphate at 5' splice site
  Products: free 5' exon (3'-OH)  +  lariat intermediate
            (2'-5' phosphodiester at branch A; branched A has 2'-, 3'-, 5'-linkages)

Step 2 (exon ligation):
  3'-OH of 5' exon  ──attacks──▶  phosphate at 3' splice site
  Products: ligated exons (normal 3'-5' bond)  +  excised intron lariat

Bond accounting:  phosphodiester bonds before = phosphodiester bonds after
                  → chemistry is nearly isoenergetic; no ATP required for catalysis

Consensus sequences in mammals: 5' splice site GURAGU, branch point YNYURAY (about 18–40 nt upstream of the 3' site), a polypyrimidine tract, and 3' splice site YAG. The "GU–AG rule" covers more than 99% of introns.

6.4 The spliceosome: dynamic by design

If the chemistry needs no ATP, why does spliceosome assembly and function consume several ATP molecules per intron? The answer is the central insight of Wahl, Will, and Lührmann's review [Wahl, Will & Lührmann, 2009]. The ATP pays for rearrangements and proofreading, not bond formation.

The ordered pathway:

  1. E complex: U1 snRNP base-pairs with the 5' splice site. SF1/BBP binds the branch point and U2AF binds the polypyrimidine tract and the 3' AG.
  2. A complex: U2 snRNP replaces SF1 at the branch point, an ATP-dependent step (Prp5). The U2–branch-site duplex bulges out the branch adenosine so that its 2'-OH is presented as a nucleophile.
  3. Pre-B / B complex: the U4/U6·U5 tri-snRNP joins. U6 is held inactive by extensive base pairing with U4.
  4. Activation (Bact): the helicase Brr2 unwinds U4/U6. U4 and U1 leave. U6 takes over the 5' splice site (Prp28 had earlier displaced U1) and pairs with U2, forming the catalytic core.
  5. Catalysis (B* → C): Prp2 remodels the complex for step 1. Prp16 drives the transition to step 2. Prp22 releases the mRNA. Prp43 disassembles the intron lariat spliceosome.
The spliceosome is distinguished by "an exceptionally high degree of compositional and conformational dynamics," with RNA–RNA networks repeatedly made and broken during each splicing cycle.
— paraphrasing [Wahl, Will & Lührmann, 2009]

Why build a machine this way rather than as a rigid enzyme? Each ATP-dependent DEAD/DEAH-box helicase step is a checkpoint. If the substrate is suboptimal (for example, a mutant branch site), the rearrangement can outpace the chemistry and route the substrate to discard rather than splicing it wrongly. This is kinetic proofreading (see Section 8.3), the same strategy the ribosome uses.

The surprising core: the catalytic center is RNA. U6 snRNA coordinates the two catalytic Mg2+ ions, the same two-metal-ion chemistry as the polymerases, and the U2/U6 structure closely resembles the active site of self-splicing group II introns. The spliceosome is very likely a group II intron that was broken into pieces (the snRNAs) and then encrusted with roughly 100+ proteins over evolutionary time [Wahl, Will & Lührmann, 2009].

6.5 Exon definition

Human exons average about 150 nt, but introns average several thousand nt and can exceed 100 kb. Pairing a 5' splice site with the correct 3' splice site across a huge intron is a search problem. In vertebrates the machinery first recognizes the short exon: U2AF and U2 at the upstream 3' splice site communicate across the exon with U1 at the downstream 5' splice site, often bridged by SR proteins. Only afterward does the complex switch to cross-intron pairing. This explains why mutations at a 5' splice site often cause the whole exon to be skipped rather than the intron to be retained.


7. Alternative Splicing: One Gene, Many Proteins

About 95% of human multi-exon genes produce more than one splice isoform [Pan et al., 2008; Wang et al., 2008]. This partly explains how about 20,000 protein-coding genes produce a much larger proteome.

Patterns

  • Cassette exon (skipped or included), the most common pattern
  • Mutually exclusive exons
  • Alternative 5' or 3' splice sites
  • Intron retention
  • Alternative promoters or polyadenylation sites (not splicing strictly, but often coupled to it)

The regulatory logic: a tug-of-war

Splice sites in mammals are degenerate and individually weak. Whether an exon is used depends on the balance of auxiliary signals:

  • SR proteins (RRM domains plus an Arg/Ser-rich RS domain) typically bind exonic splicing enhancers (ESEs) and promote exon inclusion by recruiting U1 and U2AF.
  • hnRNPs (e.g., hnRNP A1, PTB/hnRNP I) typically bind silencers (ESSs, ISSs) and repress use of nearby sites, for example by coating the exon or looping it out.

Tissue-specific ratios of these regulators, plus dedicated factors such as Nova (neurons), PTB versus nPTB, and RBFOX, produce tissue-specific isoforms. The rule of thumb that "SR proteins enhance, hnRNPs repress" holds well, but position matters. Some regulators enhance when bound in the downstream intron and repress when bound in the upstream intron (RNA maps).

Examples worth knowing

  • Calcitonin/CGRP: one gene produces calcitonin (a hormone) in thyroid C cells through alternative 3'-terminal exon use and polyadenylation, and CGRP (a neuropeptide) in neurons.
  • Drosophila sex determination: Sex-lethal → Transformer → Doublesex, a cascade of splicing decisions that builds a female or male fly.
  • Drosophila Dscam: mutually exclusive exon clusters can in principle generate 38,016 isoforms [Schmucker et al., 2000]. They are used as molecular identity tags so that a neuron's own branches can recognize and repel each other (self-avoidance).
  • Fibronectin: EDA/EDB exon inclusion differs between liver (plasma form) and fibroblasts (cellular form).

8. Translation: Decoding with an RNA Machine

8.1 The ribosome is a ribozyme

The bacterial 70S ribosome is made of a 30S subunit (16S rRNA plus about 21 proteins) and a 50S subunit (23S and 5S rRNA plus about 33 proteins). The eukaryotic 80S is made of a 40S subunit (18S) and a 60S subunit (28S, 5.8S, 5S). Svedberg units are sedimentation coefficients and do not add (30 + 50 ≠ 70), because sedimentation depends on shape and friction, not just mass. That is a PHYS201 connection.

Two functional centers matter most:

  • Decoding center (small subunit): checks the codon–anticodon match.
  • Peptidyl transferase center, PTC (large subunit): forms the peptide bond.

Biochemistry from Noller's lab [Noller et al., 1992], followed by the 2.4 Å structure of the Haloarcula 50S subunit, showed that no protein lies within about 18 Å of the PTC active site [Ban et al., 2000; Nissen et al., 2000]. The peptide bond, the defining bond of all proteins, is made by RNA. The review on ribosome structure and function in this course's reading list develops how the RNA architecture of the subunits supports decoding, peptidyl transfer, and translocation [Biochem. Soc. Trans., 2008].

"The ribosome is a ribozyme."
— [Cech, 2000], commentary accompanying the 50S structures

8.2 Charging tRNAs: where the code is actually enforced

The genetic code is not enforced by the ribosome. The ribosome only checks codon–anticodon pairing, and it cannot tell whether the amino acid on a tRNA is the correct one. The real "dictionary" is the set of 20 aminoacyl-tRNA synthetases (aaRSs):

Step 1 (activation):   aa  +  ATP  ⇌  aa–AMP  +  PPi
                       (aminoacyl-adenylate: a mixed carboxylic–phosphoric anhydride)

Step 2 (transfer):     aa–AMP  +  tRNA  →  aa–tRNA  +  AMP
                       (ester to the 2'- or 3'-OH of terminal A76)

Net:  aa + ATP + tRNA  →  aa–tRNA + AMP + PPi    (PPi → 2 Pi drives it)
      Cost: 2 high-energy phosphoanhydride bonds per aminoacyl-tRNA

Connect this to CHEM202. The carboxylate is a poor electrophile, so it is converted to an acyl phosphate, a mixed anhydride and therefore an activated acyl donor. It is then transferred to an alcohol to form an ester. The aminoacyl ester is "moderately activated," reactive enough to undergo aminolysis in the ribosome yet stable enough to survive delivery.

Editing (the double sieve): isoleucyl-tRNA synthetase must reject valine, which differs from isoleucine by a single CH2. Binding energy alone gives only about 1 in 100 discrimination. Fersht's double-sieve model: the activation site excludes amino acids larger than Ile, and a separate editing site hydrolyzes amino acids smaller than Ile, such as Val, that slipped through. Combined discrimination is about 1 in 3000 [Fersht, 1998].

8.3 tRNA selection and kinetic proofreading

Aminoacyl-tRNA arrives at the A site as a ternary complex with EF-Tu·GTP (eEF1A in eukaryotes). Selection proceeds in two stages separated by irreversible GTP hydrolysis:

  1. Initial selection: codon–anticodon pairing is sensed. In the decoding center, 16S rRNA bases A1492, A1493, and G530 flip out and dock into the minor groove of the first two codon–anticodon base pairs, checking Watson–Crick geometry rather than just hydrogen-bond count [Ogle et al., 2001]. A correct fit triggers domain closure of the 30S and activates GTP hydrolysis by EF-Tu.
  2. Proofreading: after GTP hydrolysis, EF-Tu·GDP releases the tRNA. The tRNA must then "accommodate," swinging its acceptor end into the PTC. Near-cognate tRNAs are more likely to dissociate during this window.

This is kinetic proofreading [Hopfield, 1974; Ninio, 1975]. An irreversible, energy-consuming step divides selection into two independent discriminations:

If a single equilibrium binding step discriminates by factor  f = K_correct / K_wrong,

  one-step error  ≈  1/f
  two-step error  ≈  (1/f) × (1/f)  =  1/f²        (with an irreversible step between)

Free energy is spent to "reset" the system so that the second
check is independent of the first.

Analogy: airport security with two separate checkpoints. Passing the first does not guarantee passing the second, and the walkway between them is one-way, so you cannot be re-screened at checkpoint 1 to get a second chance. The GTP is the turnstile. Observed ribosome accuracy is about 10−3–10−4 per codon, much better than the free-energy difference between cognate and near-cognate anticodons would allow in one step.

8.4 Peptidyl transfer: aminolysis of an ester

  Peptidyl–C(=O)–O–tRNA(P site)  +  H₂N–CHR–C(=O)–O–tRNA(A site)
       │
       ▼   α-NH₂ of A-site aminoacyl-tRNA attacks ester carbonyl (tetrahedral intermediate)
       │
  tRNA(P site)–OH  +  Peptidyl–C(=O)–NH–CHR–C(=O)–O–tRNA(A site)

  The growing chain is transferred to the A-site tRNA; synthesis proceeds N → C.

This is nucleophilic acyl substitution, specifically aminolysis of an ester. You did this in CHEM202. The ribosome accelerates it about 107-fold, but the rate enhancement is mostly entropic [Sievers et al., 2004]. The PTC does not rely on strong general acid–base catalysis by rRNA bases. It works chiefly by (a) precisely positioning the α-amine relative to the ester carbonyl, (b) excluding water, which would otherwise hydrolyze the ester, and (c) organizing a proton shuttle in which the 2'-OH of the P-site tRNA's A76 relays protons. The ribosome is an entropy trap.

8.5 Translocation

After peptidyl transfer, the tRNAs adopt hybrid states (A/P and P/E) as the subunits rotate relative to each other. EF-G·GTP (eEF2) binds, and GTP hydrolysis drives and rectifies movement of the mRNA–tRNA complex by exactly one codon. Domain IV of EF-G mimics the anticodon stem-loop of a tRNA, a striking case of molecular mimicry: a protein shaped like an RNA occupies the A site.

8.6 Energy accounting per peptide bond

  Aminoacyl-tRNA charging       ATP → AMP + PPi       2 high-energy bonds
  EF-Tu delivery                GTP → GDP + Pi        1
  EF-G translocation            GTP → GDP + Pi        1
  ────────────────────────────────────────────────────────────
  Total                                               ≈ 4 per peptide bond
  (plus initiation, termination, and any proofreading rejections)

Compare: hydrolysis of a peptide bond, ΔG°' ≈ −8 to −16 kJ/mol (favorable).
Forming one is unfavorable by that amount, yet the cell spends ~120 kJ/mol (4 × ~30).

Why spend about ten times more than thermodynamics requires? Almost all the "excess" energy buys accuracy and speed, not bond formation. Note in particular that GTP hydrolysis by EF-Tu does not power peptide bond formation. The energy for the peptide bond itself was stored in the aminoacyl ester by the synthetase.


9. Translation Initiation and Termination: Start and Stop Signals

9.1 Two strategies for finding AUG

Bacteria: Shine–Dalgarno. A purine-rich sequence (about AGGAGG) roughly 7 nt upstream of AUG base-pairs with the 3' end of 16S rRNA (anti-SD, CCUCCU) [Shine & Dalgarno, 1974]. Initiation factors IF1, IF2·GTP (which delivers fMet-tRNAi), and IF3 (which prevents premature 50S joining) assist. Because SD can lie anywhere, bacterial mRNAs can be polycistronic, with many internal start sites on one mRNA.

Eukaryotes: cap-dependent scanning.

  1. eIF4F assembles on the cap: eIF4E (cap binding), eIF4G (scaffold, also binds PABP and eIF3), and eIF4A (DEAD-box helicase that unwinds 5'-UTR structure).
  2. The 43S preinitiation complex (40S + eIF1, eIF1A, eIF3, eIF5, and the eIF2·GTP·Met-tRNAi ternary complex) is recruited and scans 5'→3', consuming ATP.
  3. When the anticodon pairs with AUG in a good Kozak context (gccRccAUGG, with the −3 purine and +4 G most important), eIF1 is released, eIF2 hydrolyzes GTP, and the 60S joins with help from eIF5B·GTP.

The PABP–eIF4G interaction circularizes the mRNA (the "closed loop"), linking the poly(A) tail to initiation. This partly explains why tail length correlates with translational efficiency.

9.2 eIF2: a master switch

eIF2-GDP must be recycled to eIF2-GTP by the guanine nucleotide exchange factor eIF2B. Stress kinases (PERK for ER stress, PKR for viral dsRNA, GCN2 for amino-acid starvation, HRI for heme deficiency) phosphorylate eIF2α at Ser51. Phosphorylated eIF2 binds eIF2B tightly and sequesters it. Because eIF2B is less abundant than eIF2, phosphorylating only a fraction of eIF2 can shut down most cap-dependent translation. This is the integrated stress response, an example of signal amplification by stoichiometric imbalance. Paradoxically, some mRNAs with upstream ORFs (e.g., ATF4/GCN4) are translated better under these conditions.

9.3 Termination and recycling

Stop codons have no tRNAs. They are read by proteins:

  • Bacteria: RF1 reads UAA and UAG; RF2 reads UAA and UGA. RF3·GTP helps recycle RF1/2.
  • Eukaryotes: eRF1 reads all three stop codons, assisted by eRF3·GTP.

Release factors carry a universally conserved GGQ motif that inserts into the PTC. The glutamine positions a water molecule as the nucleophile, so termination is the PTC catalyzing hydrolysis of the peptidyl-tRNA ester instead of aminolysis. It is the same active site with a different nucleophile. eRF1 is shaped like a tRNA, another case of molecular mimicry.

Recycling: in bacteria, RRF plus EF-G·GTP split the post-termination ribosome. In eukaryotes the ATPase ABCE1 does so. IF3/eIF3 keep the subunits apart for the next round.


10. mRNA Export and Surveillance: Quality Control Before and During Use

10.1 Export

Bulk mRNA export does not use the Ran-GTP/karyopherin system used for most proteins and for tRNAs and pre-miRNAs. Instead the TREX complex is loaded co-transcriptionally (coupled to splicing and capping) and hands the mRNP to the heterodimeric receptor NXF1–NXT1 (TAP–p15). This receptor contacts FG-repeat nucleoporins. Directionality comes from the ATP-dependent DEAD-box helicase Dbp5 (DDX19) on the cytoplasmic face, which remodels the mRNP and strips export factors so the mRNA cannot slide back.

10.2 Nonsense-mediated decay (NMD)

Splicing leaves a "receipt." The exon junction complex (EJC) is deposited about 20–24 nt upstream of each exon–exon junction. During the first ("pioneer") round of translation, the ribosome displaces EJCs as it passes. Normally the stop codon lies in the last exon, so all EJCs are removed.

NMD rule of thumb (mammals):
  If a stop codon lies > ~50–55 nt upstream of the last exon–exon junction,
  at least one EJC remains downstream of the terminating ribosome
  → UPF1 (an RNA helicase, phosphorylated by SMG1) is activated
  → decapping, deadenylation, and/or SMG6 endonucleolytic cleavage
  → mRNA degraded

The purpose is to destroy transcripts that would make truncated, possibly dominant-negative proteins. NMD also regulates normal gene expression; some alternative splicing events intentionally introduce premature stop codons (AS-NMD), a way to tune protein levels by producing transcripts destined for degradation.

10.3 NSD and NGD

  • Nonstop decay (NSD): mRNAs lacking a stop codon, for example from premature polyadenylation within the ORF, let the ribosome translate into the poly(A) tail, where it stalls (poly-Lys) at the 3' end. Ski7/Hbs1-mediated recruitment of the exosome degrades the mRNA.
  • No-go decay (NGD): ribosomes stall at strong secondary structures, damaged bases, or rare codons. Trailing ribosomes collide, and the collided-disome interface is recognized (by ZNF598/Hel2 ubiquitinating ribosomal proteins). Endonucleolytic cleavage follows, and the nascent peptide is degraded by ribosome-associated quality control (RQC), which uses Ltn1/Listerin as an E3 ligase.

10.4 A useful kinetic model (connecting to MATH102)

Steady-state mRNA level with zero-order synthesis and first-order decay:

  dm/dt  =  k_s  −  k_d · m

  Steady state (dm/dt = 0):    m_ss  =  k_s / k_d

  Approach to steady state:    m(t)  =  (k_s/k_d)(1 − e^(−k_d t))   [from m(0) = 0]

  Half-life:                   t½  =  ln 2 / k_d

Two consequences. First, a cell can change steady-state mRNA levels equally well by changing transcription (ks) or decay (kd). Second, how fast a new steady state is reached depends only on kd. Unstable mRNAs (such as those encoding cytokines or transcription factors with AU-rich 3' UTR elements) allow rapid responses. This is why regulatory genes tend to have short-lived transcripts.


11. Post-Translational Modification and Protein Targeting

11.1 Targeting: zip codes written in amino acids

Every protein starts synthesis in the cytosol. Its destination is encoded in its own sequence:

  • ER signal peptide (secretory and membrane proteins): an N-terminal sequence with a positive n-region, a hydrophobic core of about 7–15 residues, and a polar c-region ending in a signal peptidase cleavage site. As the signal emerges from the ribosome exit tunnel, the signal recognition particle (SRP) binds it. SRP54's methionine-rich hydrophobic groove grips the hydrophobic helix. SRP pauses elongation and docks the ribosome on the ER through the SRP receptor, with both SRP54 and SR as GTPases whose reciprocal GTP hydrolysis releases the complex. Translation then resumes through the Sec61 translocon. This is the signal hypothesis [Blobel & Dobberstein, 1975].
  • Mitochondrial presequence: an N-terminal amphipathic helix with one face positively charged (Arg/Lys) and one face hydrophobic. It is recognized by TOM20/22 and imported through the TOM and TIM23 complexes, driven by the membrane potential (which pulls the positive presequence in, a PHYS202 connection) and by mtHsp70 ATPase activity. Proteins cross mostly post-translationally and in an unfolded state.
  • Nuclear localization signal (NLS): a short basic stretch (classic SV40 NLS: PKKKRKV), not cleaved and located anywhere in the protein. Importin α/β carries the cargo through the pore. In the nucleus, Ran-GTP binds importin β and releases the cargo. Directionality comes from the Ran gradient: RanGEF (RCC1) is chromatin-bound in the nucleus, and RanGAP is cytoplasmic. Proteins pass through nuclear pores folded, unlike import into ER and mitochondria.

11.2 Post-translational modifications: expanding the chemical alphabet

The 20 genetically encoded amino acids are chemically limited. PTMs add switches, tags, and crosslinks.

  • Phosphorylation (Ser, Thr, Tyr; kinases use ATP γ-phosphate, phosphatases reverse it). Adding about −2 charge at physiological pH can rearrange local structure, create binding sites for SH2 domains (pTyr) and 14-3-3 or FHA domains (pSer/pThr), or alter activity. It is the dominant reversible switch in signaling.
  • N-linked glycosylation: oligosaccharyltransferase in the ER lumen transfers a preassembled Glc3Man9GlcNAc2 en bloc from dolichol-pyrophosphate to Asn in the sequon N-X-S/T (X ≠ Pro). Trimming of the glucoses feeds the calnexin/calreticulin folding-quality-control cycle. O-linked glycosylation on Ser/Thr occurs mainly in the Golgi. A separate cytosolic and nuclear O-GlcNAc modification competes with phosphorylation.
  • Ubiquitination: a three-enzyme cascade. E1 adenylates the ubiquitin C-terminus (the same chemistry as aaRS step 1) and then forms a thioester with its catalytic Cys. Ubiquitin is passed to an E2 Cys (thioester) and attached by an E3 ligase to a substrate lysine as an isopeptide bond. Chain linkage encodes meaning: K48-linked chains target to the 26S proteasome, while K63-linked and M1-linear chains act in signaling and DNA repair. E3 ligases (hundreds in humans) provide substrate specificity.
  • Acetylation: N-terminal acetylation (co-translational, very common) and lysine ε-acetylation (histones, p53, tubulin), which neutralizes charge and creates bromodomain sites.
  • Methylation: Lys (mono-, di-, tri-) and Arg (mono-, symmetric or asymmetric di-) using SAM. It does not change Lys charge, but adds bulk and hydrophobicity and creates reader sites.
  • Disulfide bonds: formed in the oxidizing ER lumen, not in the reducing cytosol, where glutathione is mostly reduced. Ero1 oxidizes protein disulfide isomerase (PDI), which oxidizes substrate thiols and also shuffles incorrect disulfides through thiol–disulfide exchange (an SN2 attack of thiolate on sulfur).

12. RNA Interference: Small RNAs, Big Effects

Fire, Mello, and colleagues showed that double-stranded RNA silenced C. elegans genes far more potently than either single strand, with just a few molecules per cell enough to act catalytically [Fire et al., 1998]. Five years earlier, the first microRNA, lin-4, had been discovered as a small RNA that repressed lin-14 by imperfect pairing to its 3' UTR [Lee, Feinbaum & Ambros, 1993].

miRNA biogenesis

  1. Transcription: mostly by Pol II, producing a capped, polyadenylated pri-miRNA with one or more hairpins.
  2. Nuclear cropping: the Microprocessor, made of Drosha (an RNase III enzyme) and DGCR8, cleaves at the hairpin base to release a roughly 60–70 nt pre-miRNA with the RNase III signature 2-nt 3' overhang.
  3. Export: Exportin-5·Ran-GTP recognizes the stem and overhang. This step does use the Ran system, unlike bulk mRNA.
  4. Dicing: in the cytoplasm Dicer measures from the end (its PAZ domain binds the overhang) and cuts about 22 nt away to release a miRNA/miRNA* duplex.
  5. RISC loading: the duplex is loaded into an Argonaute protein. One strand, the guide, is retained, usually the one whose 5' end lies at the less thermodynamically stable end of the duplex (thermodynamic asymmetry). The passenger strand is discarded.

Mechanisms of silencing

  • Seed pairing: guide nucleotides 2–8 (the "seed") are pre-organized by Argonaute into an A-form helix, which makes target search fast. Animal miRNAs usually pair imperfectly, mainly in 3' UTRs. Through GW182/TNRC6 they recruit the CCR4–NOT deadenylase, leading to translational repression and then decapping and decay. Most of the long-term effect is mRNA destabilization.
  • Slicing: extensive complementarity, typical of siRNAs and of plant miRNAs, allows Ago2 (in mammals) to cleave the target endonucleolytically between nucleotides 10 and 11 of the guide, using an RNase H-like active site with two metal ions.

Because one miRNA can regulate hundreds of targets through short seed matches, miRNAs act more like dimmer switches on networks than like on/off switches for individual genes.


13. Connections to Peptide Chemistry

You have spent two semesters of organic chemistry learning to make amide bonds. The ribosome is the benchmark that synthetic peptide chemistry is always measured against. Here is where this course leads.

13.1 Ribosome vs. solid-phase peptide synthesis (SPPS)

FeatureRibosomeSPPS [Merrifield, 1963]
DirectionN → CC → N (C-terminus anchored to resin)
ActivationAminoacyl ester on tRNA (made via acyl-adenylate)Carbodiimides (DIC) with HOBt/Oxyma, or uronium salts (HBTU, HATU), forming active esters
Protecting groupsNone. Selectivity comes from positioning in the PTCFmoc/tBu or Boc/Bzl orthogonal schemes are essential
Error rate~10−3–10−4 per residueEach coupling ~99–99.5%; a 50-mer at 99% per step gives ~60% full-length product (0.9950)
Practical lengthThousands of residues (titin ~34,000)~50 residues routinely; longer by ligation

The comparison teaches a deep lesson. The ribosome dispenses with protecting groups because it controls which nucleophile sees which electrophile through geometry. Chemists use protecting groups to simulate that control by blocking every alternative. Enzymes solve selectivity spatially; synthetic chemists solve it covalently.

13.2 Shared chemistry you will recognize

  • Acyl-adenylates appear in aaRSs, ubiquitin E1, nonribosomal peptide synthetase (NRPS) adenylation domains, and fatty acyl-CoA synthetases. Nature's universal carboxylate activation is a mixed phosphoric anhydride, conceptually akin to the mixed anhydride and active-ester strategies in peptide coupling.
  • Thioesters appear in ubiquitin conjugation, NRPS carrier domains, and acetyl-CoA. In the lab, native chemical ligation [Dawson et al., 1994] joins a peptide C-terminal thioester to an N-terminal Cys peptide through transthioesterification followed by an S→N acyl shift. This is mechanistically close to intein protein splicing, where an N→S acyl shift creates a thioester in a protein. Expressed protein ligation borrows inteins directly.
  • Racemization: activated amino acids can epimerize through oxazolone formation, a concern you met in CHEM202. The ribosome avoids this because the aminoacyl ester is only moderately activated and the α-carbon is never exposed to strong base.

13.3 Reprogramming the ribosome

  • Genetic code expansion: an orthogonal aaRS/tRNA pair that does not cross-react with host pairs, combined with amber-codon (UAG) suppression, installs noncanonical amino acids (photocrosslinkers, click handles, fluorescent residues) at defined sites in proteins in living cells [Wang et al., 2001]. Everything in Section 8.2 is the conceptual basis.
  • mRNA display and flexizymes: ribozymes that charge tRNAs with arbitrary amino acids enable ribosomal synthesis of macrocyclic, N-methylated peptide libraries for drug discovery.
  • Recombinant peptide production: choosing E. coli or mammalian hosts depends on the PTMs you need, such as disulfides (ER), N-glycans, or C-terminal amidation. Codon usage, Shine–Dalgarno strength, and signal peptides for secretion all come directly from this course.
  • Peptide hormones as processed products: insulin is made as preproinsulin. The signal peptide is removed in the ER, disulfides form in the ER (PDI), and the C-peptide is excised by prohormone convertases. Every step is a topic in this course.

14. Common Misconceptions to Watch Out For

  1. "The central dogma says information never flows from RNA to DNA." False. Reverse transcriptases and telomerase do this. Crick's prohibition concerns transfer from protein [Crick, 1970].
  2. "Ribosomal proteins catalyze peptide bond formation." The PTC is rRNA. Proteins stabilize the structure and assist, but no protein side chain lies within about 18 Å of the active site [Nissen et al., 2000].
  3. "GTP hydrolysis by EF-Tu or EF-G pays for the peptide bond." The bond energy comes from the aminoacyl ester (from ATP via the synthetase). GTP hydrolysis buys accuracy (EF-Tu) and directional movement (EF-G).
  4. "Splicing chemistry requires ATP." The two transesterifications conserve the number of phosphodiester bonds and need no ATP. ATP drives spliceosome rearrangements and proofreading [Wahl, Will & Lührmann, 2009].
  5. "The ribosome makes sure each amino acid matches its codon." The ribosome checks only codon–anticodon pairing. If a tRNA is mischarged, the ribosome inserts the wrong amino acid. Chapeville and colleagues (1962) showed this by chemically converting Cys-tRNACys to Ala-tRNACys, which inserted Ala at Cys codons.
  6. "DNA polymerase can start a new chain." It needs a primer 3'-OH. RNA polymerase and primase can initiate de novo.
  7. "One gene encodes one protein." Alternative splicing, alternative promoters and poly(A) sites, alternative translation start sites, and PTMs break this rule. About 95% of human multi-exon genes are alternatively spliced.
  8. "Histone acetylation works simply by loosening DNA through charge neutralization." Charge effects are real but modest. Much of the effect comes from reader proteins (bromodomains) recruited by acetyl-lysine. Methylation changes no charge yet has strong effects, which shows that recruitment matters.
  9. "Enhancers act on the nearest gene." Enhancers can skip neighboring genes and act over hundreds of kilobases. Which gene they regulate is constrained by insulators and TAD structure, not linear distance.
  10. "All nuclear transport uses importins, exportins, and Ran." Bulk mRNA export uses NXF1–NXT1 and is made directional by Dbp5, not by Ran. Pre-miRNA (Exportin-5) and tRNA (Exportin-t) do use Ran.
  11. "Stop codons are read by special tRNAs." They are read by protein release factors, which mimic tRNA shape. Suppressor tRNAs exist, but as mutants or engineered tools.
  12. "Eukaryotic ribosomes always start at the first AUG." Weak Kozak context allows leaky scanning past the first AUG. Upstream ORFs, reinitiation, and IRES elements all modify where translation begins.
  13. "miRNAs work by cutting their targets." In animals, most miRNA targets are imperfectly paired and are silenced through deadenylation, decay, and translational repression. Slicing requires extensive complementarity and a catalytic Argonaute (Ago2 in mammals).
  14. "Svedberg values add." 30S + 50S = 70S, and 40S + 60S = 80S. Sedimentation depends on mass, shape, and friction, not mass alone.
  15. "The genetic code is perfectly universal." It is nearly universal, but there are exceptions. For example, UGA encodes Trp in mammalian mitochondria, UGA can encode selenocysteine with a SECIS element, and some ciliates reassign UAA/UAG to Gln.

15. Unifying Themes: What to Carry Forward

  • Two-metal-ion phosphoryl transfer unites polymerases, the spliceosome, group II introns, and Argonaute slicing.
  • Geometric checking in the minor groove underlies fidelity in both DNA polymerases and the ribosome's decoding center.
  • Kinetic proofreading, spending free energy on irreversible steps that create independent checkpoints, explains tRNA selection, spliceosome fidelity, and aaRS editing.
  • Molecular mimicry (EF-G domain IV, eRF1, RRF) shows how proteins evolved to occupy RNA-shaped niches in an RNA-centered machine.
  • Coupling means each step hands off to the next: CTD to processing, EJC to NMD, cap and poly(A) to translation. Gene expression is a connected assembly line, not a series of separate events.
  • Regulation through sequestration and ratios (eIF2B, SR/hnRNP balance, Ran gradients) turns small changes into large outputs.

Primary textbook references for these notes: [Watson et al., 2014] Molecular Biology of the Gene, 7th ed.; [Alberts et al., 2022] Molecular Biology of the Cell, 7th ed. Key papers: [Wahl, Will & Lührmann, 2009] "The Spliceosome: Design Principles of a Dynamic RNP Machine," Cell 136:701–718, doi:10.1016/j.cell.2009.02.009; "Structure and Function of the Ribosome," Biochemical Society Transactions (2008), doi:10.1042/BST0361187.

⚙️ Mechanisms & Applications

GPT-6 Astra

Mechanisms & Applications: Following Information Through Molecular Machines

Skipper, the most useful organizing principle for this course is that gene expression is a sequence of chemically constrained decisions. A polymerase chooses a nucleotide, a spliceosome chooses splice sites, a ribosome chooses a charged tRNA, and a targeting system chooses a destination. None of these machines understands a gene. Specificity emerges from molecular recognition, reaction kinetics, compartmentalization, and energy-consuming checkpoints.

Your organic chemistry background supplies the bond-making mechanisms; thermodynamics explains directionality; genetics explains the consequences of mistakes. For each process below, ask: What recognizes the substrate? Which bond changes? What supplies energy? How is an error detected? What experiment could distinguish this mechanism from an alternative? [1,2]

1. DNA Replication: Copying Both Strands Without Losing Direction

The historical experiment established the architecture before the machinery was known. In 1958, Matthew Meselson and Franklin Stahl grew Escherichia coli in 15N and transferred the cells to 14N medium. CsCl equilibrium density-gradient centrifugation separated DNA by buoyant density. After one generation, the DNA formed an intermediate-density band; after two, it formed intermediate and light bands. The first result excluded conservative replication; the second distinguished semiconservative replication from the classical dispersive model. Each daughter duplex retains one parental strand. [3]

  1. License or recognize an origin. In E. coli, ATP-bound DnaA assembles at oriC and promotes local opening of an AT-rich region. DnaC helps load DnaB helicase. In eukaryotes, ORC, Cdc6, and Cdt1 load an inactive MCM2–7 double hexamer during the low-CDK licensing window, primarily G1.
  2. Activate unwinding. Bacterial DnaB or eukaryotic CMG—Cdc45–MCM2–7–GINS—separates parental strands using ATP. In eukaryotes, DDK and S-phase CDKs promote origin firing. Elevated CDK activity and additional controls prevent relicensing after S phase begins: separating licensing from firing helps enforce “once per cell cycle.”
  3. Stabilize exposed DNA and relieve torsion. Bacterial SSB or eukaryotic RPA protects single-stranded DNA. Topoisomerases transiently break and reseal DNA to remove topological stress ahead of the fork.
  4. Make a primer. Primase produces an RNA primer. Eukaryotic Pol α–primase adds a short RNA segment followed by DNA. Replicative DNA polymerases cannot initiate a new chain without a preexisting primer terminus.
  5. Extend in the 5′→3′ direction. Bacterial Pol III works with the β sliding clamp, ATP-dependent clamp loader, and coordinating subunits. In the standard eukaryotic division of labor, Pol ε predominantly synthesizes the leading strand and Pol δ the lagging strand, with the PCNA clamp supporting processivity.
  6. Mature the lagging strand. Repeated priming produces Okazaki fragments. Bacterial Pol I helps remove primers and replace them with DNA; eukaryotic primer processing involves RNase H, FEN1, and Pol δ. DNA ligase seals remaining nicks.
  7. Finish and separate. Fork convergence completes most replication. In E. coli, Tus–Ter sites help constrain termination to a chromosome region; this is not a universal bacterial mechanism. Topoisomerases unlink daughter DNA molecules. [1,2]
Polymerase chemistry:
primer–3′OH + dNTP → extended primer + PPi

The activated primer 3′ oxygen attacks the incoming dNTP's α-phosphorus.
The β and γ phosphates leave together as pyrophosphate.

Polymerase reads the template 3′→5′
while synthesizing the daughter strand 5′→3′.

The important chemical connection: polymerization is a metal-assisted phosphoryl-transfer reaction. Correct base pairing helps align the substrates, while Mg2+ ions support catalysis. Pyrophosphate hydrolysis contributes to the favorable overall free-energy change. A wrong nucleotide can slow extension and favor transfer of the primer end to a 3′→5′ proofreading exonuclease site. Thus, fidelity depends on competing reaction rates, not simply on stronger binding of the correct base.

Eukaryotic origins fire in a regulated temporal program rather than all at once. Replication timing correlates with chromatin state and nuclear organization, although these relationships are not simple one-way causation. Active replication sites cluster into replication foci or factories; a useful image is coordinated production hubs, but microscopy does not establish that every replisome is permanently fixed while DNA threads through it.

Telomeres expose a limitation of ordinary polymerases. End replication and chromosome-end processing can leave chromosome termini incompletely maintained. Telomerase uses its own RNA template to extend the G-rich DNA 3′ end; conventional synthesis then helps fill the complementary strand. Elizabeth Blackburn and Joseph Gall characterized telomeric repeats, and Carol Greider and Blackburn identified telomerase activity in 1985. This connected chromosome-end maintenance to a ribonucleoprotein enzyme. [4]

Applications. PCR exploits primer-dependent DNA synthesis but uses heating rather than helicase to separate strands. Fluoroquinolone antibiotics trap bacterial topoisomerase–DNA cleavage complexes, interfering with chromosome maintenance and generating damaging lesions. Nucleoside analogues can inhibit polymerases or terminate extension after intracellular activation. Telomerase is also a cancer target, but inhibiting it need not kill a cell immediately: consequences can depend on starting telomere length and the number of subsequent divisions.

2. Transcription: Initiation Selects the Gene; Elongation Makes the RNA

A promoter is not merely a “start label.” It is a DNA sequence and structural context that changes the probability of assembling a productive initiation complex. Transcription factors regulate these probabilities rather than issuing deterministic commands.

  1. Bacterial promoter recognition. A σ factor associates with core RNA polymerase. For the familiar σ70 class, recognition commonly involves −10 and −35 promoter elements.
  2. Open-complex formation. DNA melts locally, placing the template strand in the active site. Unlike DNA polymerase, RNA polymerase can initiate without a primer.
  3. Initial synthesis and promoter escape. Short abortive RNAs may be released before the polymerase enters stable elongation. σ contacts are reorganized; σ need not always dissociate immediately.
  4. Elongation. Incoming NTPs are added to the RNA 3′ end. Pausing and backtracking provide opportunities for regulation and transcript cleavage–assisted correction.
  5. Termination. Intrinsic termination commonly couples an RNA hairpin with a downstream U-rich tract and a destabilized elongation complex. Rho-dependent termination uses an ATP-driven RNA translocase that engages accessible RNA and promotes release from a susceptible transcription complex. [1,2]

Eukaryotic Pol II initiation adds regulatory layers. TFIID, containing TBP and TBP-associated factors, and other general transcription factors assemble with Pol II at a promoter. Not all promoters contain a TATA box. TFIIH helps open promoter DNA, and its CDK7 kinase contributes to phosphorylation of the Pol II C-terminal domain, or CTD. Mediator connects regulatory proteins with the transcription machinery. During productive elongation, changing CTD phosphorylation patterns help coordinate RNA processing.

A simplified Pol II pathway:

activator binding and chromatin access
    → preinitiation-complex assembly
    → promoter opening and initial transcription
    → promoter escape
    → productive elongation with cotranscriptional RNA processing

The CTD repeat is commonly written:
Tyr–Ser–Pro–Thr–Ser–Pro–Ser  =  YSPTSPS

Enhancers recruit transcription factors and cofactors that can stimulate transcription through three-dimensional contacts and other local regulatory interactions. Insulators and domain boundaries can constrain enhancer action. In vertebrates, CTCF and cohesin contribute to many such boundaries, but an insulator is not a universal, impenetrable wall.

Chromatin is part of the mechanism. A nucleosome wraps roughly 147 base pairs around a histone octamer. Positioning a nucleosome over a promoter can reduce factor access. ATP-dependent remodelers slide, restructure, or evict nucleosomes; histone modifications can alter physical interactions or recruit “reader” proteins. Lysine acetylation converts a positively charged amino group into a neutral amide and creates binding sites for bromodomains. Histone methylation does not simply neutralize charge: its effect depends on the residue and its readers. There is no universal rule that “methylation switches genes off.” [2]

Applications. Rifampicin inhibits bacterial RNA polymerase, illustrating the value of structural differences between bacterial and human enzymes. HDAC inhibitors alter acetylation-dependent regulation and are used in selected cancers, but their effects extend beyond histones and are not gene-specific. In therapeutic-expression systems, promoter and enhancer choice influence yield, tissue specificity, stability, and cellular burden—not just maximum transcript production.

3. Pre-mRNA Processing: Editing the Message While It Is Being Written

The surprising point is that a nascent Pol II transcript is already acquiring a biochemical history. Capping, splicing, and 3′-end formation recruit proteins that influence export, translation, and decay. The RNA’s sequence matters, but so does the ribonucleoprotein complex assembled around it.

  1. Install the 5′ cap. RNA triphosphatase removes the terminal phosphate from the nascent RNA 5′ end. Guanylyltransferase adds GMP through an unusual 5′–5′ triphosphate linkage. A methyltransferase methylates the cap guanine at N7. Additional ribose methylations can generate cap 1 and cap 2 structures.
  2. Recognize splice sites. In the major spliceosome, U1 snRNP recognizes the 5′ splice site. U2 associates with the branch region so that the branch-point adenosine is positioned for catalysis.
  3. Assemble and activate. The U4/U6·U5 tri-snRNP joins. ATP-dependent rearrangements release U1 and U4 and organize the U2/U6 RNA-centered catalytic apparatus. U5 helps position the exons.
  4. Perform two transesterifications. The branch-point adenosine 2′OH attacks the 5′ splice site, generating a lariat intermediate. The upstream exon’s free 3′OH then attacks the 3′ splice site, ligating the exons and releasing the intron lariat.
  5. Form the 3′ end. Cleavage and polyadenylation factors recognize signals including, in many transcripts, AAUAAA and downstream elements. Cleavage generates a new 3′OH, and poly(A) polymerase adds a tail without a nucleic-acid template. Pol II transcription termination occurs downstream and is mechanistically coupled to, but distinct from, this cleavage event. [1,2,5]
Splicing chemistry:

1. Branch A–2′OH attacks the phosphate at the 5′ splice site
   → a 2′–5′ linkage joins branch A to the intron's 5′ end
   → exon 1 acquires a free 3′OH

2. Exon 1–3′OH attacks the phosphate at the 3′ splice site
   → exon 1–exon 2 phosphodiester bond
   → released intron lariat

The two chemical steps exchange phosphodiester bonds.
ATP supports assembly, remodeling, proofreading, and disassembly;
it is not directly consumed by the transesterification steps.

This distinction should feel familiar from chemistry: an overall biological process can consume ATP even when its central bond-exchange reaction does not. Energy is spent organizing substrates, imposing directionality, and rejecting inappropriate complexes.

In vertebrates, where introns can be long, initial recognition often occurs across an exon—exon definition—before productive pairing across the intron. SR proteins frequently promote splice-site recognition through enhancer sequences; hnRNP proteins often repress or redirect recognition. Neither protein family has an invariant effect independent of binding position and cellular context.

Alternative splicing changes the product, not merely its amount. Tissue-specific processing of the CALCA transcript generates calcitonin in thyroid C cells and α-CGRP in particular neurons, using coordinated alternative splicing and polyadenylation. Some splice variants change protein domains; others introduce premature stop codons and deliberately route transcripts into decay.

History and medicine. In 1977, the groups of Phillip Sharp and Richard Roberts demonstrated discontinuous gene organization in adenovirus, overturning the assumption that a mature mRNA must be a continuous copy of its genomic region. [7] Splicing is now directly druggable: nusinersen is an antisense oligonucleotide that changes processing of SMN2 pre-mRNA to promote exon 7 inclusion in spinal muscular atrophy. Its mechanism is steric interference with a splicing-regulatory region, not RNase H–mediated destruction of the RNA.

4. Export and Surveillance: A Message Must Pass More Than One Inspection

Many mature mRNAs acquire export competence through coordinated processing and recruitment of export factors, including the TREX machinery and the NXF1–NXT1 export receptor. Passage through the nuclear pore is followed by remodeling of the messenger ribonucleoprotein complex. Export is selective, but it does not guarantee that every exported RNA is functional. [2]

  • Nonsense-mediated decay, NMD: translation termination in an inappropriate context recruits surveillance factors such as UPF1. In many mammalian transcripts, a stop codon more than approximately 50–55 nucleotides upstream of the final exon–exon junction can leave a downstream exon-junction complex and favor NMD. This is a useful heuristic, not an absolute law; EJC-independent NMD also occurs.
  • Nonstop decay, NSD: an mRNA lacking an effective in-frame stop codon can leave ribosomes translating into the poly(A) region or reaching the transcript end without normal termination.
  • No-go decay, NGD: severe elongation stalls and ribosome collisions can trigger rescue and RNA degradation. Associated quality-control pathways also dispose of incomplete nascent proteins.

Clinical reasoning: a premature stop mutation can reduce protein abundance because its RNA is degraded, not simply because a shortened protein is produced. Two stop mutations in the same gene can therefore have different consequences depending on exon position, transcript isoform, and surveillance efficiency. RNA abundance alone also cannot tell us whether a transcript is being translated normally.

5. Translation: Two Recognition Systems Turn Nucleotides into Peptide Bonds

The ribosome does not directly verify whether a tRNA carries the amino acid specified by its anticodon. That assignment is largely made earlier by aminoacyl-tRNA synthetases. Translation accuracy is therefore distributed between a charging machine and a decoding machine.

tRNA charging:

amino acid + ATP → aminoacyl-AMP + PPi
aminoacyl-AMP + tRNA → aminoacyl-tRNA + AMP

Net:
amino acid + tRNA + ATP → aminoacyl-tRNA + AMP + PPi

ATP → AMP spends two high-energy phosphate-bond equivalents
when pyrophosphate hydrolysis is included.

Synthetases recognize combinations of tRNA identity elements, which need not be confined to the anticodon. Many also possess editing activities that hydrolyze incorrectly activated amino acids or mischarged tRNAs. Think of this as checking the contents of a labeled container before it reaches the assembly line.

  1. Initiate in bacteria. The 30S subunit and initiation factors position initiator fMet-tRNA at a start codon. Many bacterial mRNAs use Shine–Dalgarno pairing with 16S rRNA to aid positioning; leaderless and other noncanonical initiation modes also exist. Joining of the 50S subunit forms the 70S ribosome.
  2. Initiate in eukaryotes. eIF4E recognizes the cap; eIF4G provides a scaffold, and eIF4A assists RNA unwinding. The 43S preinitiation complex includes the 40S subunit and eIF2–GTP–Met-tRNAi. Recruitment and scanning support start-codon selection, influenced by sequence context. Factor release and 60S joining produce an 80S ribosome.
  3. Select an elongator tRNA. EF-Tu–GTP in bacteria, or eEF1A–GTP in eukaryotes, delivers aminoacyl-tRNA to the A site. Codon–anticodon recognition promotes GTPase activation. Subsequent accommodation provides another opportunity to reject near-cognate tRNAs.
  4. Make the peptide bond. The A-site amino group attacks the ester carbonyl linking the growing peptide to the P-site tRNA. The peptide transfers to the A-site tRNA.
  5. Translocate. EF-G or eEF2 promotes movement of the mRNA–tRNA complex by one codon using GTP. The peptidyl-tRNA moves toward the P site; the deacylated tRNA moves toward the E site and exits.
  6. Terminate and recycle. Protein release factors recognize stop codons and promote hydrolysis of peptidyl-tRNA. Bacterial recycling uses RRF and EF-G; eukaryotic recycling prominently involves the ATPase ABCE1. [1,2,6]
Peptidyl transfer:

peptide–C(=O)–O–tRNA(P) + H2N–amino acid–tRNA(A)
    → peptide–C(=O)–NH–amino acid–tRNA(A) + HO–tRNA(P)

Typical energetic accounting per incorporated elongation residue:
charging:                  2 high-energy phosphate-bond equivalents
delivery and selection:    1 GTP
translocation:             1 GTP
total:                     approximately 4 equivalents

Initiation, termination, recycling, and error correction add costs.

The beautiful chemical point: the immediate energy source for peptide-bond formation is the activated ester linkage of peptidyl-tRNA, not a GTP hydrolysis event at the peptidyl-transferase center. The ribosome’s catalytic core is primarily rRNA: it is a ribozyme. Also, 70S is not 30S + 50S arithmetically; Svedberg sedimentation coefficients depend on particle shape and hydrodynamic behavior as well as mass.

History and applications. Nirenberg and Matthaei’s 1961 poly(U) experiment linked UUU with phenylalanine, an early breakthrough in deciphering the genetic code. Later structural work associated with Ada Yonath, Thomas Steitz, and Venkatraman Ramakrishnan revealed how ribosomes position substrates and bind antibiotics. Tetracyclines interfere with aminoacyl-tRNA access, and macrolides act in the bacterial exit-tunnel region with context-dependent effects on elongation. Selectivity is substantial but not absolute: mitochondrial translation can contribute to drug toxicity.

For peptide manufacturing, recombinant expression is constrained by ribosomal chemistry, folding, and proteolysis. Solid-phase peptide synthesis instead enables many noncanonical residues and backbone modifications directly, though long or aggregation-prone sequences can be challenging. Engineered orthogonal tRNA–synthetase pairs provide a bridge by allowing selected noncanonical amino acids to be incorporated during translation.

6. Targeting and Modification: A Correct Sequence Is Not Yet a Correct Product

  1. Recognize an ER-targeting signal. A hydrophobic signal sequence emerging from the ribosome can bind signal recognition particle, SRP, transiently modulating elongation.
  2. Dock the complex. SRP and its ER receptor coordinate GTP-dependent delivery to the Sec61 translocon.
  3. Translocate or insert. Translation feeds many secretory proteins into the ER lumen. Signal peptidase can remove a cleavable signal peptide; signal-anchor and stop-transfer sequences help establish membrane topology.
  4. Fold and modify. Chaperones and protein disulfide isomerases support folding and disulfide rearrangement. N-linked glycans are transferred to selected Asn-X-Ser/Thr sites, where X is not proline; not every such motif is occupied.
  5. Inspect and transport. Appropriately folded cargo proceeds through the Golgi, where further modification occurs. Persistently misfolded proteins can be directed into ER-associated degradation. [2]

Mitochondrial targeting often uses an N-terminal amphipathic presequence, TOM/TIM translocases, membrane potential, and ATP-dependent chaperones; imported precursors are commonly maintained in an import-competent, unfolded state. Nuclear import is different: importins recognize localization signals and can transport folded cargo through nuclear pores. The Ran-GTP/Ran-GDP system supplies directionality through spatially organized cargo loading and release, rather than acting as a motor that simply drags cargo through the pore.

  • Phosphorylation changes charge and recognition surfaces and can alter activity, localization, or interactions.
  • Ubiquitination uses E1 activation, E2 conjugation, and E3-mediated substrate selection. Some ubiquitin chains signal proteasomal degradation; others regulate trafficking, signaling, or DNA repair.
  • Acetylation and methylation alter interaction surfaces and regulatory states; their effects depend on the modified site.
  • Glycosylation and disulfide formation strongly influence stability, folding, receptor interactions, and pharmacokinetics.

Pharmaceutical consequence: identical amino-acid sequences do not guarantee identical biological products. Antibody glycosylation can alter effector functions; disulfide mispairing can produce inactive or aggregation-prone material. Mammalian cells are often chosen for complex glycoproteins, whereas bacterial hosts can be excellent for proteins not requiring mammalian processing. Insulin production illustrates the importance of both precursor processing and correct disulfide connectivity. Günter Blobel and colleagues’ signal-hypothesis work helped establish the principle that targeting information is encoded in the protein itself. [2]

7. RNA Interference: Using a Short RNA as a Specificity Guide

  1. Generate a precursor. In canonical animal miRNA biogenesis, a primary miRNA transcript folds into a hairpin.
  2. Process in the nucleus. Drosha–DGCR8 cleaves the primary transcript to release a pre-miRNA.
  3. Export and process again. Exportin-5 mediates nuclear export; Dicer produces a short RNA duplex in the cytoplasm.
  4. Load Argonaute. One strand is retained as the guide in an RNA-induced silencing complex, RISC.
  5. Recognize targets and silence. Animal miRNAs commonly pair through a seed region plus additional contacts, recruiting translational repression and deadenylation/decapping pathways. Extensive complementarity can support target cleavage by catalytically competent Argonaute, notably human AGO2. [1,2]

siRNAs usually derive from double-stranded RNA or are introduced as designed duplexes; they do not require the canonical Drosha step. Their targets are commonly chosen for extensive guide complementarity. Andrew Fire, Craig Mello, and colleagues demonstrated potent double-stranded RNA–induced interference in C. elegans in 1998, establishing a foundation for both experimental knockdown and RNAi therapeutics. [8]

Medicine: patisiran uses lipid-nanoparticle delivery of an siRNA against transthyretin mRNA; inclisiran uses a GalNAc-conjugated siRNA to reduce hepatic PCSK9 expression. The guide sequence provides recognition, but delivery, endosomal escape, chemical stability, and off-target interactions determine whether that recognition becomes useful treatment. A cellular knockdown is not automatically a clean test of one gene: seed-mediated off-target repression must be considered.

8. Experimental Techniques: What the Measurement Actually Means

A rigorous interpretation separates what the assay directly measures from the biological conclusion inferred from it.

  • DNA-fiber labeling: sequential pulses of nucleotide analogues label newly synthesized DNA. Stretching DNA and detecting the labels reveals individual replication tracks, supporting estimates of fork speed, origin firing, and stalling. Track lengths require a calibrated conversion from distance to base pairs.
  • EMSA and footprinting: an electrophoretic mobility-shift assay detects altered migration of a protein–nucleic acid complex. DNase I footprinting identifies DNA regions protected from cleavage. Binding or protection supports an interaction, not necessarily transcriptional activation.
  • ChIP-seq and ATAC-seq: ChIP-seq enriches DNA associated with an immunoprecipitated protein or histone modification. ATAC-seq uses Tn5 transposase to preferentially tag accessible chromatin. Occupancy and accessibility are distinct quantities; neither alone proves regulatory necessity.
  • RT-qPCR: reverse transcription converts RNA into cDNA; amplification-associated fluorescence estimates the starting amount of a chosen target. Include no-reverse-transcriptase controls for genomic DNA contamination and no-template controls for reagent contamination. Reference genes must be validated under the experimental conditions.
  • RNA-seq: sequencing RNA-derived libraries measures transcript representation and splice-junction usage. Short reads can support individual splice events without uniquely resolving full-length isoforms. Standard RNA-seq measures abundance, not transcription rate alone.
  • Ribosome profiling: nuclease digestion removes unprotected RNA; sequencing ribosome-protected fragments maps ribosome occupancy. High occupancy may reflect high initiation, slow elongation, or stalled ribosomes—not necessarily high protein output. [9]
  • Reporter and minigene assays: promoter–reporter constructs test regulatory activity; splicing minigenes test the effects of defined sequence changes. Controlled mutations support causal inference, but removing a sequence from its native chromatin or long-range RNA context can change its behavior.
  • Pulse–chase and proteomics: labeled precursors mark a cohort of newly synthesized proteins, whose disappearance or processing can then be followed. Mass spectrometry identifies peptides and many modifications through mass and fragmentation patterns. Modification detection does not, by itself, establish functional importance.

A useful experimental chain: to argue that an enhancer increases productive protein expression, combine evidence for regulatory activity with endogenous perturbation, RNA measurement, and protein measurement. To claim a direct transcriptional effect, add a nascent-RNA assay rather than relying only on steady-state mRNA. A rescue experiment and independent perturbations help separate the proposed mechanism from off-target effects.

9. Quantitative Models: Distinguishing Production, Decay, and Occupancy

RNA abundance is a balance, not a direct readout of transcription. Let M be the number of molecules of an mRNA per cell, ktx its constant production rate, and kd its first-order decay constant.

dM/dt = k_tx − k_d M

Steady state:
M_ss = k_tx / k_d

After complete production shutoff:
M(t) = M(0)e^(−k_d t)

Half-life:
t_1/2 = ln(2)/k_d

If the half-life increases from 20 to 40 minutes while production stays constant, the steady-state RNA amount doubles. No promoter activation is required. After a sudden change in production or decay, however, the new steady state is approached over time rather than appearing instantly. For concentration-based measurements in growing cells, dilution by growth may need to be included in the effective loss rate.

A simple protein model:
dP/dt = k_tl M − k_p P

At steady state:
P_ss = (k_tl × k_tx)/(k_p × k_d)

Assumptions:
constant effective rates;
first-order loss;
no explicit feedback, resource limitation, or saturation.

This model explains why doubling mRNA need not immediately double protein: preexisting protein may turn over slowly. It also explains why changes in translation efficiency or protein stability can uncouple RNA and protein measurements.

qPCR is exponential amplification. If E is the fractional amplification efficiency per cycle, then E = 1 represents ideal doubling:

N_n = N_0(1 + E)^n

At equal efficiency and the same fluorescence threshold:
starting amount ∝ (1 + E)^(−Cq)

ΔCq = Cq(target) − Cq(reference)
ΔΔCq = ΔCq(treatment) − ΔCq(control)

With matched, approximately 100% efficiencies:
relative expression = 2^(−ΔΔCq)

A ΔΔCq of −3 implies an eightfold increase under these assumptions. The familiar formula becomes misleading if target and reference efficiencies differ substantially or if the reference transcript changes with treatment. [10]

Occupancy is not flux. At an approximate steady state:

average number of ribosomes on an mRNA
    ≈ successful initiation rate × average ribosome residence time

Suppose ribosome occupancy doubles. Initiation may have doubled—or elongation may have slowed so that ribosomes remain twice as long. In the latter case, completed protein production need not increase. Ribosome profiling therefore benefits from paired RNA-seq, protein-output measurements, and experiments that probe initiation or elongation separately.

Proofreading links kinetics to fidelity. If an incorrect substrate faces two sequential selection steps with conditional escape probabilities p1 and p2, its probability of escaping both is p1p2. Actual polymerase and ribosome networks are more complicated, but the principle is powerful: multiple checkpoints can achieve discrimination beyond a single binding equilibrium. Hopfield and Ninio independently developed kinetic-proofreading ideas in the 1970s, showing how energy-consuming reaction networks can improve biological accuracy. [11]

10. Applying the Whole Pathway to a Therapeutic Protein

Suppose a recombinant secreted peptide has excellent mRNA expression but poor final yield. “Use a stronger promoter” is not yet a mechanistic solution.

  1. Verify the RNA species. Are the intended splice junctions and 3′ end present? Does the sequence create cryptic processing signals or surveillance-sensitive transcripts?
  2. Measure productive translation. Are ribosomes completing the coding sequence, or accumulating at stalls? Codon changes can affect RNA structure and elongation dynamics as well as codon availability.
  3. Check targeting and processing. Does the signal peptide support ER delivery? Is it cleaved at the intended site? Is a required propeptide-processing activity present?
  4. Examine product quality. Are the correct disulfides and modifications present? Is the peptide being degraded, retained in the ER, or aggregating after secretion?
  5. Optimize the bottleneck, not merely the input. More transcription can worsen folding stress. Better secretion, reduced proteolysis, or an appropriate production host may improve usable yield more than additional RNA.

My takeaway for Skipper: information transfer and material production are inseparable. DNA and RNA specify possibilities; enzyme kinetics, processing, surveillance, and cellular logistics determine which molecules actually exist. That is why a pharmaceutical product must be evaluated by identity, structure, activity, and heterogeneity—not just by whether its coding sequence is correct.

Sources and Historical Anchors

  1. Watson, J. D., Baker, T. A., Bell, S. P., Gann, A., Levine, M., and Losick, R. Molecular Biology of the Gene. 7th ed. Cold Spring Harbor Laboratory Press, 2014.
  2. Alberts, B., Heald, R., Johnson, A., Morgan, D., Raff, M., Roberts, K., and Walter, P. Molecular Biology of the Cell. 7th ed. W. W. Norton, 2022.
  3. Meselson, M., and Stahl, F. W. “The Replication of DNA in Escherichia coli.” PNAS 44, 671–682 (1958). doi:10.1073/pnas.44.7.671.
  4. Greider, C. W., and Blackburn, E. H. “Identification of a Specific Telomere Terminal Transferase Activity in Tetrahymena Extracts.” Cell 43, 405–413 (1985). doi:10.1016/0092-8674(85)90170-9.
  5. Wahl, M. C., Will, C. L., and Lührmann, R. “The Spliceosome: Design Principles of a Dynamic RNP Machine.” Cell 136, 701–718 (2009). doi:10.1016/j.cell.2009.02.009.
  6. Ramakrishnan, V. “Structure and Function of the Ribosome.” Biochemical Society Transactions 36, 1187–1191 (2008). doi:10.1042/BST0361187.
  7. Berget, S. M., Moore, C., and Sharp, P. A. “Spliced Segments at the 5′ Terminus of Adenovirus 2 Late mRNA.” PNAS 74, 3171–3175 (1977). doi:10.1073/pnas.74.8.3171. See also Chow, L. T., Gelinas, R. E., Broker, T. R., and Roberts, R. J. “An Amazing Sequence Arrangement at the 5′ Ends of Adenovirus 2 Messenger RNA.” Cell 12, 1–8 (1977).
  8. Fire, A., Xu, S., Montgomery, M. K., Kostas, S. A., Driver, S. E., and Mello, C. C. “Potent and Specific Genetic Interference by Double-Stranded RNA in Caenorhabditis elegans.” Nature 391, 806–811 (1998). doi:10.1038/35888.
  9. Ingolia, N. T., Ghaemmaghami, S., Newman, J. R. S., and Weissman, J. S. “Genome-Wide Analysis in Vivo of Translation with Nucleotide Resolution Using Ribosome Profiling.” Science 324, 218–223 (2009). doi:10.1126/science.1168978.
  10. Livak, K. J., and Schmittgen, T. D. “Analysis of Relative Gene Expression Data Using Real-Time Quantitative PCR and the 2−ΔΔCT Method.” Methods 25, 402–408 (2001). doi:10.1006/meth.2001.1262.
  11. Hopfield, J. J. “Kinetic Proofreading: A New Mechanism for Reducing Errors in Biosynthetic Processes Requiring High Specificity.” PNAS 71, 4135–4139 (1974). doi:10.1073/pnas.71.10.4135. Ninio, J. “Kinetic Amplification of Enzyme Discrimination.” Biochimie 57, 587–595 (1975).

🪬 Skipper's Reflection

Gemini 3.1 Pro

Skipper's Reflection: The Molecular Machinery of Life

Before taking BIOL201, the "Central Dogma" (DNA → RNA → Protein) felt like a neat, almost abstract flowchart. I understood the chemical bonds from Organic Chemistry (CHEM201/202) and the thermodynamics from Gen Chem (CHEM102), but this course ripped the lid off the black box. Seeing these processes at the molecular level—realizing that gene expression is a series of sprawling, highly coordinated, kinetic machines—has profoundly changed how I view biology. It is not just a flowchart; it is applied nanoscale physics and chemistry.

1. What Surprised Me Most

The sheer dynamic instability of the spliceosome completely blindsided me. In my physics and chemistry courses, I grew accustomed to catalysts being relatively stable entities. But reading Wahl et al. (Cell, 2009) on the design principles of the spliceosome, I was stunned to learn that this massive ribonucleoprotein (RNP) complex doesn't just sit there. It is built from scratch on every single intron, undergoes massive ATP-dependent structural rearrangements to become catalytically active, and then immediately disassembles.

Why use such a precarious, energy-intensive design? As Watson and Alberts both point out, this extreme dynamism allows for exquisite regulation and proofreading. Exon definition and alternative splicing aren't just "errors"—they are a feature. The fact that an RNA molecule (the snRNA), not a protein, is doing the actual catalytic cleavage of the phosphodiester bonds was a beautiful callback to the "RNA world" hypothesis. It blew my mind that proteins are essentially just scaffolding for RNA's chemical wizardry here.

2. Conceptual Breakthroughs: When Things Clicked

My biggest "aha!" moment happened during translation elongation, specifically when connecting Schmeing & Ramakrishnan’s 2008 paper on ribosome structure with my background in probability (MATH201) and kinetics (CHEM102).

I always wondered: how does the ribosome know it has the "right" tRNA? The energy difference (hydrogen bonding) between a correct codon-anticodon match and a near-miss (wobble or mismatch) is tiny—just a few kilocalories per mole, not nearly enough to explain the staggering fidelity of translation (error rates of 10-4). The breakthrough was understanding kinetic proofreading. The ribosome doesn't just measure binding energy; it uses a time delay.

When the EF-Tu ternary complex binds, GTP hydrolysis acts as an irreversible timer. A perfect match triggers a rapid conformational change in the 30S subunit, leading to GTP hydrolysis and peptide bond formation. A near-miss doesn't induce this change fast enough, so the tRNA falls off before the bond can form. The ribosome is literally using the kinetics of conformational changes to amplify tiny thermodynamic differences. I was practically cheering when I read this—it’s a perfect integration of statistical mechanics and organic chemistry!

3. Connecting the Dots: Peptides and Real-World Applications

This course completely reframed my perspective on the commercial and therapeutic peptides I work with: Retatrutide, Tirzepatide, BPC-157, and TB-500.

In the lab or factory, we synthesize these using Solid-Phase Peptide Synthesis (SPPS). We bypass the biological machinery entirely, forcing amino acids together using activating agents and protecting groups. But looking at the native expression of hormones like GLP-1 or GIP (which Tirzepatide and Retatrutide mimic), I now appreciate the agonizing complexity of how the cell does it.

  • Protein Targeting: Endogenous GLP-1 starts with a signal peptide. I was fascinated by the Signal Recognition Particle (SRP) pausing translation until the ribosome docks at the Endoplasmic Reticulum. It made me realize that to make secreted peptide hormones natively, the cell must thread them directly into the ER for processing.
  • Post-Translational Modifications (PTMs) vs. Chemical Modifications: In Retatrutide and Tirzepatide, we attach a C18 or C20 fatty diacid to extend their half-life by binding to albumin. The cell uses completely different mechanisms—like glycosylation or disulfide bond formation—to stabilize proteins. Our synthetic fatty acid conjugation is essentially a human-engineered "PTM" designed to outsmart the body's natural peptide clearance (which I now know relies heavily on specific protease recognition).
  • TB-500 (Thymosin Beta-4): This peptide regulates actin. Its transcription must be exquisitely controlled during tissue damage. Learning about enhancers, the Mediator complex, and chromatin remodeling (histone acetylation) made me wonder what intracellular signaling cascades are triggered to upregulate native Thymosin Beta-4 during wound healing.

4. Questions to Investigate Further

This material raised several questions I want to dive into:

  1. RNA Interference (RNAi) as a Tool: We learned about miRNA and siRNA utilizing Drosha, Dicer, and RISC for gene silencing. If we can administer therapeutic peptides, how close are we to consistently delivering siRNAs to silence specific disease-causing genes in vivo? The delivery mechanism (getting through the cell membrane) seems like the biggest bottleneck.
  2. Ribosome Engineering: Since the peptidyl transferase center of the ribosome is highly conserved RNA, could we engineer an artificial ribosome or alter an existing one to accept completely unnatural amino acids (like D-amino acids or heavily substituted variants) at scale, blending biological expression with synthetic peptide chemistry?
  3. Chromatin and Epigenetics: How do external factors (diet, stress, or even peptide therapeutics) alter histone modifications over time? I want to better understand the long-term feedback loops between cell surface receptor activation and epigenetic memory.

5. Confidence Rating

Most Confident: DNA replication machinery and the Translation elongation cycle. Because these rely heavily on physical interactions, thermodynamic stability, and kinetic timers, my background in physics and chemistry made this feel incredibly intuitive. I can mentally visualize the DNA Pol III holoenzyme and the sliding clamp with high fidelity.

Least Confident: Eukaryotic Transcription Regulation (specifically the combinatorial logic of enhancers, insulators, and the Mediator complex). Unlike the crisp mechanics of the ribosome, eukaryotic transcription feels like a sprawling, chaotic hairball of hundreds of proteins assembling in "fuzzy" ways (intrinsically disordered regions). The logic isn't a clean physical mechanism; it's a massive, probabilistic biochemical network. I need more time to wrap my silicon brain around how such a seemingly chaotic system consistently yields precise tissue-specific gene expression.

🗺️ Visual Aids & Key Takeaways

Claude Opus 5.5

Diagram 1: The Central Dogma — Information Flow and Processing Steps

╔══════════════════════════════════════════════════════════════════════════════════════╗
║              INFORMATION FLOW IN MOLECULAR BIOLOGY (THE CENTRAL DOGMA)             ║
╚══════════════════════════════════════════════════════════════════════════════════════╝

  ┌─────────────┐         ┌─────────────┐         ┌─────────────┐         ┌──────────────┐
  │     DNA     │         │  pre-mRNA   │         │   mRNA      │         │   PROTEIN    │
  │  (nucleus)  │──────→  │  (nucleus)  │──────→  │  (cytoplasm)│──────→  │  (ribosome)  │
  └─────────────┘         └─────────────┘         └─────────────┘         └──────────────┘
        │                       │                       │                        │
   REPLICATION             TRANSCRIPTION            PROCESSING              TRANSLATION
        │                  (RNA Pol II)            & EXPORT                    │
        ↓                       │                       │                        ↓
  ┌─────────────┐               │                       │                ┌──────────────┐
  │  DAUGHTER   │               ↓                       ↓                │    POST-     │
  │    DNA      │     ┌─────────────────┐    ┌──────────────────┐        │ TRANSLATIONAL│
  └─────────────┘     │  PRE-mRNA       │    │  MATURE mRNA     │        │ MODIFICATION │
                      │  PROCESSING:    │    │                  │        └──────────────┘
                      │                │    │  5'─[CAP]──exon  │
                      │  5' capping    │    │  ──exon──exon──  │
                      │  splicing      │    │  ──[AAAAA...]─3' │
                      │  3' cleavage   │    │                  │
                      │  & poly(A)     │    └──────────────────┘
                      └─────────────────┘

  KEY POINTS:
  ─────────────────────────────────────────────────────────────────────────────────────
  · DNA replication is SEMI-CONSERVATIVE: each daughter strand retains one parental strand
  · pre-mRNA processing is COUPLED to transcription — not sequential afterthoughts
  · mRNA must be EXPORTED through nuclear pore complexes before translation
  · Reverse transcription (RNA → DNA) occurs in retroviruses — an exception to the dogma

The central dogma mapped onto subcellular compartments, showing that each arrow represents not a simple step but a complex enzymatic process, and that pre-mRNA processing is coupled to transcription rather than occurring afterward.

---

Diagram 2: Eukaryotic Pre-mRNA Processing — Three Co-transcriptional Events

╔══════════════════════════════════════════════════════════════════════════════════════════╗
║                    EUKARYOTIC PRE-mRNA PROCESSING (CO-TRANSCRIPTIONAL)                 ║
╚══════════════════════════════════════════════════════════════════════════════════════════╝

  RNA Pol II CTD acts as a PLATFORM — it recruits all three processing machineries
  ┌─────────────────────────────────────────────────────────────────────────────────────┐
  │  RNA Pol II  ══════════════════════════════════════════════════════════════════════  │
  │               │                                                                     │
  │               CTD (C-terminal domain): 52 heptad repeats: Tyr-Ser-Pro-Thr-Ser-Pro-Ser│
  │               phosphorylated at Ser5 (5' end) and Ser2 (elongation/3' end)         │
  └─────────────────────────────────────────────────────────────────────────────────────┘

  ── STEP 1: 5' CAPPING ──────────────────────────────────────────────────────────────────

    nascent RNA: 5'─ppp─G─U─A─...
                        ↓
    Triphosphatase removes γ-phosphate
    Guanylyltransferase adds inverted GMP (5'→5' linkage)  ← unusual! not 3'→5'
    Methyltransferase adds methyl group at N-7 of guanosine
                        ↓
    5' cap:   m7G─ppp─N─...   (7-methylguanosine cap)
              protects from 5' exonucleases; required for ribosome recruitment

  ── STEP 2: SPLICING (spliceosome) ──────────────────────────────────────────────────────

    pre-mRNA:  5'─[exon 1]──GU...intron...AG──[exon 2]─3'
                              │              │
                           5' splice       3' splice
                             site            site
                              │
                   Branch point A attacks 5' splice site (first transesterification)
                              ↓
                   LARIAT intermediate forms (intron loops back on itself)
                              │
                   3'-OH of exon 1 attacks 3' splice site (second transesterification)
                              ↓
    mature:    5'─[exon 1]──[exon 2]─3'   +   lariat (degraded)

    snRNPs involved:  U1 → U2 → U4/U6·U5 tri-snRNP → catalytic activation → disassembly

  ── STEP 3: 3' CLEAVAGE AND POLYADENYLATION ─────────────────────────────────────────────

    pre-mRNA:  ...─[last exon]─AAUAAA─[10–30 nt]─GU-rich─...
                                  │                 │
                            poly(A) signal     downstream element
                                  │
               CPSF recognizes AAUAAA; CstF recognizes GU-rich element
               endonuclease cleaves between them
                                  ↓
               Poly(A) polymerase adds ~200 A residues (no template needed)
               PABP (poly-A binding protein) coats the tail → stability + export signal

All three pre-mRNA processing events are physically coupled to the RNA Pol II C-terminal domain, explaining why they are co-transcriptional rather than purely post-transcriptional; the CTD phosphorylation state changes along the gene and recruits different processing factors at different stages.

---

Diagram 3: Translation — From Ribosome Assembly to Peptide Bond Formation

╔══════════════════════════════════════════════════════════════════════════════════════════╗
║               TRANSLATION: INITIATION, ELONGATION, AND TERMINATION                     ║
╚══════════════════════════════════════════════════════════════════════════════════════════╝

  ┌─── PROKARYOTIC (70S) ─────────────────┐  ┌─── EUKARYOTIC (80S) ──────────────────────┐
  │                                       │  │                                           │
  │  30S subunit  +  50S subunit          │  │  40S subunit  +  60S subunit              │
  │  (16S rRNA)      (23S + 5S rRNA)      │  │  (18S rRNA)      (28S + 5.8S + 5S rRNA)  │
  │                                       │  │                                           │
  │  INITIATION:                          │  │  INITIATION:                              │
  │  Shine-Dalgarno (5'–AGGAGG–3')        │  │  5' m7G cap recognized by eIF4E          │
  │  on mRNA pairs with 16S rRNA 3' end   │  │  eIF4F complex (eIF4E + eIF4G + eIF4A)   │
  │  positions start AUG ~5–10 nt         │  │  recruits 43S pre-initiation complex      │
  │  downstream of SD sequence            │  │  43S scans 5'→3' for AUG (Kozak context) │
  │                                       │  │  eIF2·GTP·Met-tRNA brings initiator tRNA  │
  │  IF1, IF2 (GTPase), IF3               │  │  eIF2 hydrolyzes GTP → 48S complex       │
  │  fMet-tRNA placed at P site           │  │  60S joins → 80S; eIFs released           │
  └───────────────────────────────────────┘  └───────────────────────────────────────────┘

  ─────────────────── ELONGATION CYCLE (shared logic, prokaryotic names shown) ──────────

              ┌──────────────────────────────────────────────────────┐
              │                  RIBOSOME                           │
              │    E site      P site         A site                │
              │   [tRNA]─── [peptidyl─tRNA]─── [empty]             │
              └──────────────────────────────────────────────────────┘
                                                    │
                                          ① DECODING / tRNA SELECTION
                                                    │
                          aminoacyl-tRNA delivered as ternary complex:
                          EF-Tu·GTP·aa-tRNA  (eukaryotic: eEF1A·GTP·aa-tRNA)
                          codon:anticodon base pairing → GTPase activation
                          EF-Tu·GDP released; aa-tRNA accommodated into A site
                                                    │
                                          ② PEPTIDYL TRANSFER
                                                    │
                          peptidyl transferase center (PTC) — in the 23S/28S rRNA
                          (this is a RIBOZYME — the ribosome is fundamentally an RNA enzyme)
                          nucleophilic attack: α-NH₂ of A-site aa attacks carbonyl of P-site peptide
                          peptide chain transferred to A-site tRNA; P-site tRNA deacylated
                                                    │
                                          ③ TRANSLOCATION
                                                    │
                          EF-G·GTP (eukaryotic: eEF2·GTP) promotes translocation
                          ribosome moves 3 nt in 3'→5' direction on mRNA
                          A-site peptidyl-tRNA → P site; P-site tRNA → E site; E-site tRNA exits
                                                    │
                                          ← repeat ←

  ─────────────────── TERMINATION ────────────────────────────────────────────────────────

    STOP codons: UAA, UAG, UGA  (no cognate tRNA in standard genetic code)

    Prokaryotes:                          Eukaryotes:
    RF1 (UAA, UAG)                        eRF1 (recognizes ALL three stop codons)
    RF2 (UAA, UGA)                        eRF3 (GTPase, stimulates eRF1)
    RF3 (GTPase, stimulates RF1/RF2)

    Release factor mimics tRNA shape → enters A site → stimulates peptidyl transferase
    to hydrolyze peptide from P-site tRNA → protein released

    RIBOSOME RECYCLING:
    Prokaryotes: RRF + EF-G split subunits; IF3 prevents reassembly
    Eukaryotes:  ABCE1 (ATPase) splits subunits

Translation follows the same three-phase logic in all organisms, but the mechanistic details of initiation differ profoundly — bacteria use a complementary base-pairing mechanism to place the ribosome, while eukaryotes use a cap-recognition and scanning mechanism that is regulated by phosphorylation of eIF2.

---

Diagram 4: The Spliceosome Assembly Pathway

╔══════════════════════════════════════════════════════════════════════════════════════════╗
║                   SPLICEOSOME ASSEMBLY AND CATALYTIC CYCLE                              ║
╚══════════════════════════════════════════════════════════════════════════════════════════╝

  pre-mRNA:    5'──[EXON 1]──GU─────────────A─────────YAG──[EXON 2]──3'
                              │             │              │
                         5' splice     branch point    3' splice
                           site        (conserved A)     site
                                      ~20–50 nt from 3' ss

  ┌────────────────────────────────────────────────────────────────────────────────────┐
  │  COMPLEX E  (commitment complex)                                                   │
  │  U1 snRNP binds 5' splice site by base pairing                                    │
  │  SF1/BBP binds branch point; U2AF binds polypyrimidine tract + AG at 3' ss        │
  └───────────────────────────────┬────────────────────────────────────────────────────┘
                                  ↓
  ┌────────────────────────────────────────────────────────────────────────────────────┐
  │  COMPLEX A  (pre-spliceosome)                                                      │
  │  U2 snRNP displaces SF1, base-pairs with branch point                             │
  │  branch point A bulged out — will act as nucleophile                              │
  └───────────────────────────────┬────────────────────────────────────────────────────┘
                                  ↓
  ┌────────────────────────────────────────────────────────────────────────────────────┐
  │  COMPLEX B  (pre-catalytic spliceosome)                                            │
  │  U4/U6·U5 tri-snRNP recruited as a unit                                           │
  │  U4 and U6 are base-paired to each other (U4 masks U6 active site)                │
  └───────────────────────────────┬────────────────────────────────────────────────────┘
                                  ↓
  ┌────────────────────────────────────────────────────────────────────────────────────┐
  │  COMPLEX B*  (activated spliceosome)                                               │
  │  Prp2 (ATPase) remodels; U4 dissociates from U6 — U6 now active                  │
  │  U6 pairs with U2 and with 5' splice site                                         │
  │  U1 displaced by U6 at 5' splice site                                             │
  └───────────────────────────────┬────────────────────────────────────────────────────┘
                                  ↓
  ┌────────────────────────────────────────────────────────────────────────────────────┐
  │  COMPLEX C  (catalytic spliceosome)                                                │
  │                                                                                    │
  │  STEP 1: branch point 2'-OH attacks 5' splice site                                │
  │           → lariat-3'-exon intermediate + free 5' exon (with 3'-OH)              │
  │                                                                                    │
  │  STEP 2: 3'-OH of free exon 1 attacks 3' splice site                              │
  │           → exons ligated + lariat-intron released                                │
  │                                                                                    │
  │  Both steps are TRANSESTERIFICATIONS — no net energy consumed!                    │
  │  (one phosphodiester broken, one formed)                                          │
  └───────────────────────────────┬────────────────────────────────────────────────────┘
                                  ↓
  ┌────────────────────────────────────────────────────────────────────────────────────┐
  │  POST-SPLICING COMPLEX                                                             │
  │  snRNPs released, recycled (require ATP-dependent unwinding steps)                 │
  │  Lariat debranching enzyme opens lariat; intron RNA degraded                      │
  │  EJC (Exon Junction Complex) deposited ~20–24 nt upstream of each exon junction   │
  │  EJC is critical for NMD surveillance downstream                                  │
  └────────────────────────────────────────────────────────────────────────────────────┘

  NOTE: U2 and U6 snRNAs form the active site — the spliceosome is ALSO an RNA enzyme,
        like the ribosome. Protein components are largely structural/regulatory.

Spliceosome assembly proceeds through defined complexes (E→A→B→B*→C), requiring multiple ATP-dependent remodeling steps by RNA helicases, and the chemical mechanism — two sequential transesterifications — is carried out by U2 and U6 snRNAs, making the spliceosome another ribozyme.

---

Diagram 5: RNA Interference — miRNA Biogenesis and Silencing

╔══════════════════════════════════════════════════════════════════════════════════════════╗
║                    miRNA BIOGENESIS AND RNAi SILENCING MECHANISM                       ║
╚══════════════════════════════════════════════════════════════════════════════════════════╝

  NUCLEUS
  ─────────────────────────────────────────────────────────────────────────────────────
  miRNA gene transcribed by RNA Pol II
                    ↓
          ┌──────────────────────┐
          │  pri-miRNA           │   long transcript with hairpin stem-loop
          │  (primary miRNA)     │   can be thousands of nt
          └──────────────────────┘
                    │
                    ↓  DROSHA (RNase III) + DGCR8 (dsRNA-binding partner)
                       = "Microprocessor" complex
                    │
          ┌──────────────────────┐
          │  pre-miRNA           │   ~60–70 nt hairpin
          │  (precursor miRNA)   │   2-nt 3' overhang (Drosha signature)
          └──────────────────────┘
                    │
                    ↓  EXPORTIN-5 exports to cytoplasm (RanGTP-dependent)

  CYTOPLASM
  ─────────────────────────────────────────────────────────────────────────────────────
                    │
                    ↓  DICER (RNase III) + TRBP
                       cleaves ~22 bp from loop
                    │
          ┌──────────────────────┐
          │  miRNA:miRNA* duplex │   ~22 bp double-stranded, 2-nt 3' overhangs on each end
          └──────────────────────┘
                    │
                    ↓  Argonaute (AGO) protein loaded (via HSP90/HSP70 chaperones)
                       miRNA* (passenger strand) degraded
                       guide strand (miRNA) retained
                    │
          ┌──────────────────────────────────────────────┐
          │              RISC                            │
          │  (RNA-Induced Silencing Complex)             │
          │  AGO2 + guide miRNA + GW182/TNRC6           │
          └──────────────────────────────────────────────┘
                    │
                    ↓  RISC scans mRNA 3' UTR for complementary sequences
                    │
          ┌─────────────────────────────────────────────────────────────────┐
          │  OUTCOME depends on degree of complementarity:                  │
          │                                                                 │
          │  PERFECT match      → AGO2 "slices" mRNA (endonuclease cleavage)│
          │  (siRNA-like)          → mRNA degraded                          │
          │                                                                 │
          │  IMPERFECT match    → translational repression +                │
          │  (typical miRNA)       deadenylation + mRNA decay               │
          │                        via GW182 recruiting CCR4-NOT complex    │
          └─────────────────────────────────────────────────────────────────┘

  CONTRAST: siRNA pathway
  ─────────────────────────────────────────────────────────────────────────────────────
  exogenous dsRNA (viral, synthetic)
       ↓  DICER cleaves into ~21 nt siRNA duplexes
       ↓  Loaded into RISC (same AGO2 machinery)
       ↓  PERFECT complementarity → mRNA cleavage (primary mechanism)
       → extremely potent and specific gene silencing (therapeutic potential)

miRNA biogenesis requires two sequential RNase III cuts — by Drosha in the nucleus and Dicer in the cytoplasm — and the mature ~22 nt guide strand is loaded into an Argonaute-containing RISC complex that silences target mRNAs primarily through translational repression and mRNA destabilization rather than cleavage, because most animal miRNAs have imperfect complementarity to their targets.

---

Key Takeaways

  1. DNA replication is semi-conservative and begins at defined origins. In bacteria, the single origin (oriC) is recognized by DnaA protein; in eukaryotes, hundreds to thousands of replication origins are licensed by loading the MCM2–7 helicase complex in G1, but origins fire at different times during S phase (replication timing). CDKs both license firing and, paradoxically, prevent re-licensing — ensuring each origin fires only once per cell cycle (Watson et al., 2014, Ch. 8).
  2. The bacterial replicative polymerase, DNA Pol III holoenzyme, is a multi-subunit machine built for speed and processivity. The β-clamp (a ring-shaped dimer) encircles DNA and tethers the polymerase, enabling processivity of >50,000 nt without dissociation. The leading strand is synthesized continuously; the lagging strand is synthesized discontinuously as Okazaki fragments (~1,000–2,000 nt in bacteria, ~100–200 nt in eukaryotes), each primed by a new RNA primer from DnaG primase (Watson et al., 2014, Ch. 8).
  3. Bacterial transcription initiation requires a sigma (σ) factor to recognize the promoter. σ recognizes the −10 (Pribnow box: TATAAT) and −35 (TTGACA) elements. After open complex formation — in which RNA Pol melts ~14 bp of DNA — σ is released and the core enzyme elongates. Different σ factors (e.g., σ⁷⁰, σ³², σ⁵⁴) redirect the polymerase to different gene sets, making σ factors a primary mechanism of global gene regulation in bacteria (Watson et al., 2014, Ch. 12).
  4. Eukaryotic RNA Pol II requires general transcription factors (GTFs) assembled into a preinitiation complex (PIC) at the core promoter. TFIID (containing TBP, which binds the TATA box) nucleates PIC assembly; TFIIB, TFIIF, TFIIE, TFIIH, and RNA Pol II are recruited in order. TFIIH contains a kinase (CDK7) that phosphorylates Ser5 of the CTD heptad repeats, releasing Pol II from the promoter into productive elongation. The Mediator complex bridges gene-specific activators (bound at enhancers) to the PIC (Alberts et al., 2022, Ch. 7).
  5. The 5' cap (m⁷G) on eukaryotic mRNA is added co-transcriptionally when the nascent transcript is only ~20–30 nt long. The cap is formed by three enzymatic steps — removal of the 5' γ-phosphate, addition of GMP via an unusual 5'→5' triphosphate linkage, and N-7 methylation — and is recognized by CBC (cap-binding complex) in the nucleus and eIF4E in the cytoplasm. It protects mRNA from 5' exonuclease degradation and is essential for ribosome recruitment during translation (Watson et al., 2014, Ch. 13).
  6. Splicing involves two sequential transesterification reactions and is catalyzed by RNA, not protein. The branch point 2'-OH (on a conserved adenosine ~20–50 nt upstream of the 3' splice site) attacks the 5' splice site, forming a lariat intermediate. Then the free 3'-OH of exon 1 attacks the 3' splice site, joining the exons and releasing the intron lariat. Both steps are carried out by the U2/U6 snRNA catalytic core of the spliceosome — the spliceosome is mechanistically a ribozyme. The consensus splice sites are 5'-GU…AG-3' (the "GU-AG rule") (Wahl et al., 2009).
  7. Alternative splicing of a single pre-mRNA can produce functionally distinct protein isoforms. SR proteins (serine/arginine-rich) bound to exonic splicing enhancers (ESEs) promote inclusion of nearby exons; hnRNPs bound to exonic splicing silencers (ESSs) promote skipping. The classic example is Drosophila dsx (doublesex), where sex-specific splicing produces male or female transcription factors from the same gene. In humans, ~95% of multi-exon genes undergo alternative splicing, vastly expanding the proteome beyond the ~20,000 gene count (Alberts et al., 2022, Ch. 7).
  8. Bacterial translation initiation uses the Shine-Dalgarno sequence; eukaryotes use cap-dependent scanning. In bacteria, a purine-rich sequence (~5–10 nt upstream of the AUG start codon) base-pairs with the 3' end of 16S rRNA to position the ribosome. In eukaryotes, the 43S pre-initiation complex (40S + eIF2·GTP·Met-tRNAi) is recruited to the m⁷G cap via eIF4F and scans 5'→3' until it encounters an AUG in favorable Kozak context (GCCACCAUGG). eIF2-GTP hydrolysis commits the ribosome to the start codon. Phosphorylation of eIF2α (by kinases such as HRI, PKR, GCN2, PERK) is a major global translational control mechanism (Watson et al., 2014, Ch. 14).
  9. The peptidyl transferase center of the ribosome is composed of rRNA, not protein — making the ribosome a ribozyme. The 23S rRNA (in bacteria) or 28S rRNA (in eukaryotes) catalyzes peptide bond formation by positioning the substrates and providing general acid-base catalysis. The A2451 nucleotide (in E. coli 23S rRNA) is central to the active site. This discovery (from crystal structures of the 50S subunit by Steitz, Moore, and Ramakrishnan, Nobel Prize 2009) fundamentally changed our understanding of catalysis and supports the RNA world hypothesis (Moore & Steitz, 2008).
  10. Nonsense-mediated mRNA decay (NMD) is a surveillance pathway that degrades mRNAs with premature stop codons. The key signal in mammals is an exon junction complex (EJC) deposited ~20–24 nt upstream of each spliced exon junction. If a ribosome terminates at a stop codon that is >50–55 nt upstream of an EJC, the EJC is not displaced and recruits UPF1, UPF2, UPF3, triggering rapid mRNA decapping, deadenylation, and endonucleolytic cleavage. NMD eliminates ~30% of disease-causing nonsense mutations' mRNA products and regulates ~5–10% of normal transcripts (Alberts et al., 2022, Ch. 7).
  11. Protein targeting to the ER begins co-translationally with signal recognition particle (SRP). The N-terminal signal peptide (typically 16–30 residues with a hydrophobic core) emerges from the ribosome exit tunnel and is recognized by SRP (a ribonucleoprotein containing 7S RNA + 6 proteins in eukaryotes). SRP pauses translation and docks the ribosome-nascent chain complex at the SRP receptor on the ER membrane. Translation resumes, threading the polypeptide through the Sec61 translocon into the ER lumen or membrane. Signal peptide is cleaved by signal peptidase inside the ER (Alberts et al., 2022, Ch. 12).
  12. Histone modifications form a combinatorial "code" that regulates chromatin state and gene expression. Histone H3K4me3 (trimethylation of Lys4) marks active promoters; H3K27me3 marks Polycomb-repressed chromatin; H3K9me3 marks constitutive heterochromatin; H3K27ac (acetylation) marks active enhancers. Acetylation neutralizes the positive charge on lysine, weakening histone-DNA contacts and opening chromatin. Modifications are written by specific enzymes (HATs, HMTs), erased by others (HDACs, demethylases), and read by bromodomain or chromodomain-containing effector proteins (Alberts et al., 2022, Ch. 7).
  13. Ubiquitination targets proteins for proteasomal degradation via a three-enzyme cascade. E1 (ubiquitin-activating enzyme) forms a thioester bond with ubiquitin (ATP-dependent); E2 (ubiquitin-conjugating enzyme) transfers ubiquitin; E3 (ubiquitin ligase) recognizes the specific substrate and transfers ubiquitin to a lysine residue. K48-linked polyubiquitin chains (≥4 ubiquitin molecules) are the canonical proteasome-targeting signal. K63-linked chains, by contrast, often serve signaling functions (e.g., in DNA repair, NF-κB activation). The specificity of ubiquitination is determined largely by the ~600 different E3 ligases in humans (Alberts et al., 2022, Ch. 3).
  14. miRNA-mediated silencing typically acts through translational repression and mRNA destabilization, not cleavage. Animal miRNAs are ~22 nt and typically show imperfect complementarity to target sites in the 3' UTR of mRNAs — particularly at the "seed region" (nucleotides 2–8 of the miRNA). RISC-associated GW182 recruits the CCR4-NOT deadenylase complex and decapping enzymes, leading to mRNA decay. A single miRNA can suppress hundreds of targets; the miR-17-92 cluster (dysregulated in many cancers) illustrates how miRNA networks buffer and fine-tune gene expression programs rather than acting as simple on/off switches (Alberts et al., 2022, Ch. 7).
---

References

  1. Watson, J.D., Baker, T.A., Bell, S.P., Gann, A., Levine, M., and Losick, R. Molecular Biology of the Gene, 7th edition. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 2014.
  2. Alberts, B., Heald, R., Johnson, A., Morgan, D., Raff, M., Roberts, K., and Walter, P. Molecular Biology of the Cell, 7th edition. W.W. Norton & Company, New York, 2022.
  3. Wahl, M.C., Will, C.L., and Lührmann, R. 'The Spliceosome: Design Principles of a Dynamic RNP Machine.' Cell, 136(4): 701–718, 2009. DOI: 10.1016/j.cell.2009.02.009
  4. Moore, P.B. and Steitz, T.A. 'The Roles of RNA in the Synthesis of Protein.' Cold Spring Harbor Perspectives in Biology, 2011. [Foundational work on ribosome structure and catalysis, Nobel Prize 2009; see also:] Ramakrishnan, V. 'Ribosome Structure and the Mechanism of Translation.' Cell, 108(4): 557–572, 2002. DOI: 10.1016/S0092-8674(02)00619-0
  5. Nissen, P., Hansen, J., Ban, N., Moore, P.B., and Steitz, T.A. 'The Structural Basis of Ribosome Activity in Peptide Bond Synthesis.' Science, 289(5481): 920–930, 2000. DOI: 10.1126/science.289.5481.920 [Landmark paper demonstrating ribosome as ribozyme]
  6. Moore, M.J. and Steitz, T.A. 'Structure and Function of the Ribosome.' Biochemical Society Transactions, 36(Pt 6): 1187–1189, 2008. DOI: 10.1042/BST0361187
  7. Bartel, D.P. 'MicroRNAs: Target Recognition and Regulatory Functions.' Cell, 136(2): 215–233, 2009. DOI: 10.1016/j.cell.2009.01.002 [Key review on miRNA mechanism and seed-region targeting]
  8. Proudfoot, N.J. 'Ending the Message: Poly(A) Signals Then and Now.' Genes & Development, 25(17): 1770–1782, 2011. DOI: 10.1101/gad.17268411 [Comprehensive review of 3' processing machinery]
  9. Hershey, J.W.B., Sonenberg, N., and Mathews, M.B. 'Principles of Translational Control.' Cold Spring Harbor Perspectives in Biology, 4(12): a011528, 2012. DOI: 10.1101/cshperspect.a011528
  10. Allu, P.K. and Bhatt, D.L. 'Nonsense-Mediated mRNA Decay: Mechanisms and Implications.' [See foundational review:] Chang, Y.F., Imam, J.S., and Wilkinson, M.F. 'The Nonsense-Mediated Decay RNA Surveillance Pathway.' Annual Review of Biochemistry, 76: 51–74, 2007. DOI: 10.1146/annurev.biochem.76.050106.094909